Bacterial Persistence: Methods and Protocols (Methods in Molecular Biology, 2357) [2nd ed. 2021] 107161620X, 9781071616208

Table of contents :
Preface
Contents
Contributors
Part I: Introduction
Chapter 1: Studying Bacterial Persistence: Established Methods and Current Advances
1 Bacterial Persistence
2 Clinical Relevance
3 Challenges in Persistence Research
4 Current Methods
4.1 Genetic Screenings
4.2 Experimental Evolution
4.3 Transcriptomics and Proteomics
4.3.1 Transcriptomics
4.3.2 Proteomics
4.4 Single-cell Techniques
4.4.1 Single-cell Time-Lapse Microscopy
4.4.2 Flow Cytometry
4.4.3 Advances in Single-cell Technologies
4.5 Mathematical Modelling
4.6 In vivo Models
5 Future Perspectives
References
Part II: Quantification and Enrichment of Persister Cells
Chapter 2: Antibiotic Tolerance and Persistence Studied Throughout Bacterial Growth Phases
1 Introduction
2 Materials
3 Methods
3.1 Determination of Antibiotic Tolerance Along the Growth Curve
3.1.1 Day 0: Collection of Materials and Inoculation of the Overnight Culture
3.1.2 Day 1: Subculturing
3.1.3 Day 2: Antibiotic Tolerance Assay
3.1.4 Day 3: CFU Counting
3.1.5 Day 4: CFU Counting and Plotting
3.2 Determination of Antibiotic Tolerance in Stationary Phase
3.2.1 Day 0: Preparation of Media and Cultures
3.2.2 Day 1: Subculturing
3.2.3 Day 3: Stationary-Phase Antibiotic Tolerance Assay
3.2.4 Day 4: CFU Counting
3.2.5 Day 5: CFU Counting and Plotting
4 Notes
References
Chapter 3: A Robust Method for Generating, Quantifying, and Testing Large Numbers of Escherichia coli Persisters
1 Introduction
2 Materials
2.1 Generation of Large Fractions of Tolerant Cells
2.2 Identification and Quantification of Tolerant Cells Using Membrane Staining and Flow Cytometry
2.3 Assessment of Antibiotic Tolerance with Flow Cytometry
3 Methods
3.1 Generation of Large Fractions of Tolerant Cells
3.2 Identification and Quantification of Tolerant Cells Using Membrane Staining and Flow Cytometry
3.3 Guide on How to Use the Matlab Script
3.4 Assessment of Antibiotic Tolerance with Flow Cytometry
4 Notes
References
Chapter 4: Enrichment of Persister Cells Through -Lactam-Induced Filamentation and Size Separation
1 Introduction
2 Materials
3 Methods
3.1 Installation of the Vacuum Filtration System
3.2 Persister Enrichment Protocol
4 Notes
References
Chapter 5: Methods for Enrichment of Bacterial Persister Populations for Phenotypic Screens and Genomic Studies
1 Introduction
2 Materials
3 Methods
3.1 Preparation of Exponentially Dividing Bacterial Cultures
3.2 Determination of Minimal Bactericidal Antibiotic Concentration
3.3 Enrichment Method #1: Flow Sorting for Persisters by Staining Membrane Depolarization
3.4 Enrichment Method #2: Antibiotic Lysis of Nonpersister Cells
3.5 Enrichment Method #3: Culture Aging to Induce Persister Cell Survival
3.6 Bacterial Cell Lysis with Lysozyme Pretreatment
3.7 Bacterial RNA Extraction
4 Notes
References
Part III: Single-Cell Analysis of Persister Cells
Chapter 6: Observing Bacterial Persistence at Single-Cell Resolution
1 Introduction
2 Materials
2.1 Determining Time Delay in Killing of Growing Cells
2.2 Single-Cell Observation of Lag Time After Starvation
3 Methods
3.1 Determining Time Delay in Killing of Growing Cells
3.2 Single-Cell Observation of Lag Time After Starvation
3.3 Preparing the Agarose Pad
3.4 Sandwiching Cells
3.5 Microscopy
3.6 MicrobeJ Analysis
4 Notes
References
Chapter 7: Phenotypic Characterization of Antibiotic Persisters at the Single-Cell Level: From Data Acquisition to Data Analys...
1 Introduction
2 Materials
3 Methods
3.1 Data Acquisition
3.1.1 Day 1: Preparation of Cultures and Medium
3.1.2 Day 2: Precultures
3.1.3 Day 3: Cultures and Microscopy
3.2 Data Analysis
3.2.1 Cell Segmentation, Cell Tracking and Intensity/Morphology Measurements
3.2.2 Defining and Monitoring Key Parameters for Persistence
3.2.3 Persister Identification
3.2.4 Microcolony Growth Rate
3.2.5 Single-Cell Elongation Rate
3.2.6 Persisters Growth Stage Identification
3.2.7 Fluorescent Biosensors
4 Notes
References
Chapter 8: Microfluidics for Single-Cell Study of Antibiotic Tolerance and Persistence Induced by Nutrient Limitation
1 Introduction
2 Materials
2.1 Bacterial Strains and Growth Media
2.2 Preparation of Microfluidics Device
2.3 Cell Observation and Microfluidic Experiments
2.4 Additional Equipment for Setup with Temporal Heterogeneity in Nutrient Availability
3 Methods
3.1 Preparation of Microfluidics Devices
3.2 Setup with Spatial Heterogeneity in Nutrient Availability
3.3 Setup with Temporal Heterogeneity in Nutrient Availability
4 Notes
References
Chapter 9: Counting Chromosomes in Individual Bacteria to Quantify Their Impacts on Persistence
1 Introduction
2 Materials
2.1 Bacterial Strains and Plasmids
2.2 Media
2.3 Nucleic Acid Staining
2.4 Persister Assay
2.5 Fluorescence Activated Cell Sorting (FACS)
2.6 Microscopy
3 Methods
3.1 Cell Growth Conditions
3.2 Nucleic Acid Staining Selection
3.2.1 Hoechst 33342 Staining Procedure
3.2.2 Determination of the Impact of Hoechst 33342 Staining on Culturability and Persistence
3.3 Sample Preparation for Fluorescence Activated Cell Sorting (FACS)
3.3.1 Cell Dilution and Staining for FACS
3.3.2 Other Preparations for FACS
3.4 Controls for FACS: Determination of Whether the Timing Required to Sort Cells Using FACS Influences Persistence
3.5 FACS
3.6 Verification of Sorting Fidelity Based on DNA Content
3.6.1 Confirmation of DNA Content of Sorted Cells Using a Secondary dsDNA-Specific Dye (PicoGreen)
3.6.2 Confirmation of DNA Content of Sorted Cells Using an Origin-of-Replication Reporter
3.6.3 Controls for Origin Reporter
3.6.4 Confocal Microscopy
3.7 Culturability and Persistence Assays on Cell Samples
3.7.1 Culturability and Persistence Assay on Sorted Fractions
3.7.2 Culturability and Persistence Assays on Unsorted Cells Diluted to the Same Density of Sorted Cells
4 Notes
References
Part IV: Omics Approaches in Persistence Research
Chapter 10: Analyzing Persister Proteomes with SILAC and Label-Free Methods
1 Introduction
2 Materials
2.1 Bacterial Strains
2.2 Media and Reagents
2.2.1 Protein Sample Preparation
3 Methods
3.1 Pulsed-SILAC of Persister Cells
3.2 Persister Recovery
3.3 Sample Preparation for MS
3.4 MS Analysis
4 Notes
References
Chapter 11: Defining Proteomic Signatures to Predict Multidrug Persistence in Pseudomonas aeruginosa
1 Introduction
2 Materials
2.1 General Materials
2.2 Whole Cell Lysis and Protein Extraction
2.3 Sample Preparation for LC-MS/MS: Protein Digestion
2.4 Sample Preparation for LC-MS/MS: Solid-Phase Peptide Purification
2.5 LC-MS/MS
3 Methods
3.1 Sample Generation
3.2 Whole Cell Lysis and Protein Extraction
3.3 Sample Preparation for LC-MS/MS: Protein Digestion
3.4 Sample Preparation for LC-MS/MS: Solid-Phase Peptide Purification
3.5 LC-MS/MS Analysis
3.6 Peptide Level Data Set Generation
3.7 Regression and Feature Reduction of Proteomic Models of Persistence
3.7.1 Dataset Preparation
3.7.2 Orthogonal Partial Least Square Regression (oplsr) of MS Data
4 Notes
References
Chapter 12: The Use of Experimental Evolution to Study the Response of Pseudomonas aeruginosa to Single or Double Antibiotic T...
1 Introduction
2 Materials
2.1 General Materials
2.2 In Vitro Evolution Experiment
2.3 Determination of MIC
2.4 Genomic DNA Preparation
2.5 Isolation of Hypertolerant or Drug Resistant Strains
3 Methods
3.1 Evolution Experiment
3.2 MIC Determination
3.3 Preparation of gDNA and Genome Sequencing
3.3.1 Alignment to a Reference Genome
3.3.2 Call, Selection, and Quantification of Emerging Alleles
3.3.3 Identification of Regions of Recombination
3.4 Isolation of Individual Highly Tolerant or Drug Resistant Strains
3.5 Analysis of the Antibiotic-Related Characteristics of Individual Isolates
3.6 The Contribution of Low-Level Resistance to Survival and Tolerance
4 Notes
References
Part V: Persister Cell Resuscitation and Eradication
Chapter 13: Detecting Persister Awakening Determinants
1 Introduction
2 Materials
3 Methods
3.1 Identification of Persister Awakening Genes by FACS
3.1.1 Persister Assay and Control Samples
3.1.2 Preparatory Work
3.1.3 Sample Preparation
3.1.4 Sorting and Plating Metabolically Active Cells
3.1.5 Data Analysis
3.2 Identification of Persister Awakening Genes by Flow Cytometry
3.2.1 Persister Assay and Control Samples
3.2.2 Preparatory Work
3.2.3 Sample Preparation
3.2.4 Measuring the Number of Metabolically Active Cells
3.2.5 Plating Out the Samples
3.2.6 Data Analysis
3.3 Validation by Single-Cell Growth Dynamics
3.3.1 Sample Preparation
3.3.2 Visualization of Regrowth Using Time-Lapse Microscopy
3.3.3 Data Analysis
4 Notes
References
Chapter 14: Monitoring Persister Resuscitation with Flow Cytometry
1 Introduction
2 Materials
3 Methods
3.1 Preparing Overnight Precultures
3.2 Setting Up the Flow Cytometry Parameters
3.3 Beta-Lactam Mediated Cell Lysis
3.4 Clonogenic Survival Assay
3.5 Monitoring Persister Resuscitation
4 Notes
References
Chapter 15: Stimulating Aminoglycoside Uptake to Kill Staphylococcus aureus Persisters
1 Introduction
2 Materials
3 Methods
3.1 Quantifying Persister Cells
3.2 Determining Adjuvant Concentration
3.3 Evaluating Killing Against Chemically Induced Tolerant S. aureus
3.4 Texas Red-Conjugated Tobramycin Uptake
4 Notes
References
Part VI: Cellular and Animal Model Systems for Studying Persistence
Chapter 16: In Vitro Models for the Study of the Intracellular Activity of Antibiotics
1 Introduction
2 Materials
2.1 Equipment
2.2 Reagents
3 Methods
3.1 Preparation of Bacterial Suspension and of Media
3.2 Opsonization of Bacteria
3.3 Preparation of Eukaryotic Cells and Bacteria for Infection
3.4 Phagocytosis
3.5 Intracellular Growth
3.6 Assessment of Antibiotic Intracellular Activity
4 Notes
References
Chapter 17: Analysis of Salmonella Persister Population Sizes, Dynamics of Gut Luminal Seeding, and Plasmid Transfer in Mouse ...
1 Introduction
2 Materials
3 Methods
3.1 Identification of Adequate Donor-Recipient Pairs
3.1.1 Determining the MIC of Relevant Strains to Antibiotics Used in the Mouse Model
3.1.2 In Vitro Conjugation Tests Between Salmonella Donors and Enterobacteriaceae Recipients to Characterize Plasmids of Inter...
3.1.3 Identifying a Suitable Donor Strain in an Animal Model for Typhoidal Salmonellosis
3.1.4 Identifying a Suitable Recipient Strain in an Animal Model
3.1.5 Tagging of Plasmids with WITS Tags Coupled to Antibiotic Resistance Markers
3.2 Infection of Mice with the Donor and Antibiotic Treatment Schedule to Yield Tissue Lodged Persisters
3.2.1 Generation of Mixed WITS Inoculum
3.2.2 Animal Infection and Antibiotic Treatment Schedule
3.2.3 Euthanasia of Mice and Analysis of Persister Reservoirs and Population Sizes
3.2.4 Enrichment of S.Tm WITS from Fecal or Organ Homogenates
3.3 Introduce Recipient Strains by Oral Gavage and Analysis of Transconjugant Formation in the Gut Lumen
3.3.1 Infecting Mice Harboring Tissue-Lodged Persisters with Recipient Strains
3.3.2 Euthanasia of Mice at the End of the Experiment, Analysis of Conjugation Dynamics, and the Formation of Reservoirs of S....
3.4 Quantification of WITS Tag Abundance by qPCR
3.4.1 qPCR Analysis of WITS Tag Abundance
3.4.2 Analysis of the Population Structure and Detection of Bottlenecks or Rare Events
3.5 Modifications to the Protocol for Alternate Applications: Oral Infection with S.Tm Donor Strains as an Example
4 Notes
References
Chapter 18: Studying Antibiotic Persistence During Infection
1 Introduction
2 Materials
2.1 Quantification of Antibiotic Persisters In Vitro and In Vivo
2.2 Infection of Bone-Marrow Macrophages
2.3 Visualization of Heterogeneity in a Bacterial Population During Infection
2.4 Bacterial Persistence Using DualRNAseq
3 Methods
3.1 Studying Antibiotic Persistence In Vitro (Fig. 1a)
3.2 Studying Antibiotic Persistence In Vivo
3.2.1 Infection of Bone-Marrow Macrophages (BMM)
3.2.2 Quantification of Antibiotic Persisters In Vivo (Fig. 1b)
3.2.3 Tracking and Quantifying Growing and Non-growing Bacteria (Fig. 2)
3.2.4 Distinction Between Active and Inactive Non-growers (Fig. 3a)
3.2.5 Overview of Bacterial Persistence Using Dual RNAseq
4 Notes
References
Correction to: Studying Antibiotic Persistence During Infection
Index

Citation preview

Methods in Molecular Biology 2357

Natalie Verstraeten Jan Michiels Editors

Bacterial Persistence Methods and Protocols Second Edition

METHODS

IN

MOLECULAR BIOLOGY

Series Editor John M. Walker School of Life and Medical Sciences University of Hertfordshire Hatfield, Hertfordshire, UK

For further volumes: http://www.springer.com/series/7651

For over 35 years, biological scientists have come to rely on the research protocols and methodologies in the critically acclaimed Methods in Molecular Biology series. The series was the first to introduce the step-by-step protocols approach that has become the standard in all biomedical protocol publishing. Each protocol is provided in readily-reproducible step-bystep fashion, opening with an introductory overview, a list of the materials and reagents needed to complete the experiment, and followed by a detailed procedure that is supported with a helpful notes section offering tips and tricks of the trade as well as troubleshooting advice. These hallmark features were introduced by series editor Dr. John Walker and constitute the key ingredient in each and every volume of the Methods in Molecular Biology series. Tested and trusted, comprehensive and reliable, all protocols from the series are indexed in PubMed.

Bacterial Persistence Methods and Protocols Second Edition

Edited by

Natalie Verstraeten and Jan Michiels VIB-KU Leuven Center for Microbiology, VIB, Leuven, Belgium; Centre of Microbial and Plant Genetics, KU Leuven, Leuven, Belgium

Editors Natalie Verstraeten VIB-KU Leuven Center for Microbiology VIB Leuven, Belgium

Jan Michiels VIB-KU Leuven Center for Microbiology VIB Leuven, Belgium

Centre of Microbial and Plant Genetics Centre of Microbial and Plant Genetics KU Leuven Leuven, Belgium KU Leuven Leuven, Belgium

ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-0716-1620-8 ISBN 978-1-0716-1621-5 (eBook) https://doi.org/10.1007/978-1-0716-1621-5 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2021, Corrected Publication 2022 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer Nature. The registered company address is: 1 New York Plaza, New York, NY 10004, U.S.A.

Preface At the time of this writing, we are in the midst of a global pandemic that presents unprecedented challenges to health systems. Countless lives are lost, and the economic, social, and psychological impact of COVID-19 can hardly be underestimated. This clearly illustrates the devastating power microorganisms can have over human beings. In the case of bacteria, antibiotics have for many years provided an invaluable tool to clear infections and to facilitate surgical procedures and therapies that require immune suppression, including cancer treatments. However, bacteria are developing strategies to circumvent the lethal action of antibiotics at an alarmingly fast pace. In this respect, not only the well-known problem of resistance but also bacterial persistence comes to light. Persister cells comprise a small, transiently multidrug tolerant subpopulation in an otherwise susceptible group of bacteria. These cells contribute to the chronic character of many infections and play a role in the development of genetic resistance, warranting the renewed scientific interest in persistence since the 1980s. However, the transient nature of persister cells and their scarce numbers present a need for dedicated techniques and protocols. This volume combines state-of-theart methods that are currently used in persistence research and can be instrumental in further deciphering this intriguing phenomenon. Leuven, Belgium

Natalie Verstraeten Jan Michiels

v

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

PART I

INTRODUCTION

1 Studying Bacterial Persistence: Established Methods and Current Advances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Elen Louwagie, Laure Verstraete, Jan Michiels, and Natalie Verstraeten

PART II

3

QUANTIFICATION AND ENRICHMENT OF PERSISTER CELLS

2 Antibiotic Tolerance and Persistence Studied Throughout Bacterial Growth Phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Enea Maffei, Cinzia Fino, and Alexander Harms 3 A Robust Method for Generating, Quantifying, and Testing Large Numbers of Escherichia coli Persisters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Silke R. Vedelaar, Jakub L. Radzikowski, and Matthias Heinemann 4 Enrichment of Persister Cells Through B-Lactam-Induced Filamentation and Size Separation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Etthel Windels, Bram Van den Bergh, and Jan Michiels 5 Methods for Enrichment of Bacterial Persister Populations for Phenotypic Screens and Genomic Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Samantha Adikari, Elizabeth Hong-Geller, and Sofiya Micheva-Viteva

PART III

v ix

23

41

63

71

SINGLE-CELL ANALYSIS OF PERSISTER CELLS

6 Observing Bacterial Persistence at Single-Cell Resolution . . . . . . . . . . . . . . . . . . . . 85 Emma Dawson, Emrah S¸ims¸ek, and Minsu Kim 7 Phenotypic Characterization of Antibiotic Persisters at the Single-Cell Level: From Data Acquisition to Data Analysis . . . . . . . . . . . . . . . . . . . 95 Nathan Fraikin, Laurence Van Melderen, and Fre´de´ric Goormaghtigh 8 Microfluidics for Single-Cell Study of Antibiotic Tolerance and Persistence Induced by Nutrient Limitation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Stefany Moreno-Ga´mez, Alma Dal Co, Simon van Vliet, and Martin Ackermann 9 Counting Chromosomes in Individual Bacteria to Quantify Their Impacts on Persistence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125 Allison M. Murawski, Katherine Rittenbach, Christina J. DeCoste, Gary Laevsky, and Mark P. Brynildsen

vii

viii

Contents

PART IV 10 11

12

Analyzing Persister Proteomes with SILAC and Label-Free Methods. . . . . . . . . . 149 Bork A. Berghoff Defining Proteomic Signatures to Predict Multidrug Persistence in Pseudomonas aeruginosa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 Pablo Manfredi, Isabella Santi, Enea Maffei, Emmanuelle Lezan, Alexander Schmidt, and Urs Jenal The Use of Experimental Evolution to Study the Response of Pseudomonas aeruginosa to Single or Double Antibiotic Treatment . . . . . . . . . 177 Isabella Santi, Pablo Manfredi, and Urs Jenal

PART V 13 14 15

OMICS APPROACHES IN PERSISTENCE RESEARCH

PERSISTER CELL RESUSCITATION AND ERADICATION

Detecting Persister Awakening Determinants. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 Dorien Wilmaerts, Jan Michiels, and Natalie Verstraeten Monitoring Persister Resuscitation with Flow Cytometry . . . . . . . . . . . . . . . . . . . . 209 Sayed Golam Mohiuddin and Mehmet A. Orman Stimulating Aminoglycoside Uptake to Kill Staphylococcus aureus Persisters . . . . 223 Ashelyn E. Sidders, Lauren C. Radlinski, Sarah E. Rowe, and Brian P. Conlon

PART VI

CELLULAR AND ANIMAL MODEL SYSTEMS FOR STUDYING PERSISTENCE

16

In Vitro Models for the Study of the Intracellular Activity of Antibiotics. . . . . . . 239 Fre´de´ric Peyrusson, Tiep K. Nguyen, Julien M. Buyck, Sandrine Lemaire, Gang Wang, Cristina Seral, Paul M. Tulkens, and Franc¸oise Van Bambeke 17 Analysis of Salmonella Persister Population Sizes, Dynamics of Gut Luminal Seeding, and Plasmid Transfer in Mouse Models of Salmonellosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 253 Erik Bakkeren, Joshua P. M. Newson, and Wolf-Dietrich Hardt 18 Studying Antibiotic Persistence During Infection . . . . . . . . . . . . . . . . . . . . . . . . . . . 273 Charlotte Michaux, Se´verin Ronneau, and Sophie Helaine Correction to: Studying Antibiotic Persistence During Infection . . . . . . . . . . . . . . . . . . C1 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

291

Contributors MARTIN ACKERMANN • Department of Environmental Systems Science, Institute of Biogeochemistry and Pollutant Dynamics, ETH Zurich, Zurich, Switzerland; Department of Environmental Microbiology, Eawag, Du¨bendorf, Switzerland SAMANTHA ADIKARI • Bioscience Division, Los Alamos National Laboratory, Los Alamos, NM, USA ERIK BAKKEREN • Department of Biology, Institute of Microbiology, ETH Zurich, Zu¨rich, Switzerland BORK A. BERGHOFF • Institute for Microbiology and Molecular Biology, Justus Liebig University Giessen, Giessen, Germany MARK P. BRYNILDSEN • Department of Molecular Biology, Princeton University, Princeton, NJ, USA; Department of Chemical and Biological Engineering, Princeton University, Princeton, NJ, USA JULIEN M. BUYCK • Pharmacologie Cellulaire et Mole´culaire, Louvain Drug Research Institute, Universite´ catholique de Louvain, Brussels, Belgium; INSERM U1070 “Pharmacology of Anti-infective Agents”, Universite´ de Poitiers, Poitiers, France BRIAN P. CONLON • Department of Microbiology and Immunology, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA; Marsico Lung Institute, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA ALMA DAL CO • School of Engineering and Applied Sciences, Harvard University, Cambridge, MA, USA EMMA DAWSON • Department of Physics, Emory University, Atlanta, GA, USA CHRISTINA J. DECOSTE • Department of Molecular Biology, Princeton University, Princeton, NJ, USA CINZIA FINO • Focal Area Infection Biology, Biozentrum, University of Basel, Basel, Switzerland; Department of Biology, University of Copenhagen, Copenhagen, Denmark NATHAN FRAIKIN • Cellular and Molecular Microbiology, Faculte´ des Sciences, Universite´ Libre de Bruxelles (ULB), Gosselies, Belgium FRE´DE´RIC GOORMAGHTIGH • Biozentrum, University of Basel, Basel, Switzerland WOLF-DIETRICH HARDT • Department of Biology, Institute of Microbiology, ETH Zurich, Zu¨rich, Switzerland ALEXANDER HARMS • Focal Area Infection Biology, Biozentrum, University of Basel, Basel, Switzerland MATTHIAS HEINEMANN • Molecular Systems Biology, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, Groningen, The Netherlands SOPHIE HELAINE • Section of Microbiology, Medical Research Council Centre for Molecular Bacteriology and Infection, Imperial College London, London, UK; Department of Microbiology, Harvard Medical School, Boston, MA, USA ELIZABETH HONG-GELLER • Bioscience Division, Los Alamos National Laboratory, Los Alamos, NM, USA URS JENAL • Biozentrum, University of Basel, Basel, Switzerland MINSU KIM • Department of Physics, Emory University, Atlanta, GA, USA; Graduate Division of Biological and Biomedical Sciences, Emory University, Atlanta, GA, USA

ix

x

Contributors

GARY LAEVSKY • Department of Molecular Biology, Princeton University, Princeton, NJ, USA SANDRINE LEMAIRE • Pharmacologie Cellulaire et Mole´culaire, Louvain Drug Research Institute, Universite´ catholique de Louvain, Brussels, Belgium; GSK Biologicals, Rixensart, Belgium EMMANUELLE LEZAN • Biozentrum, University of Basel, Basel, Switzerland ELEN LOUWAGIE • VIB-KU Leuven Center for Microbiology, VIB, Leuven, Belgium; Centre of Microbial and Plant Genetics, KU Leuven, Leuven, Belgium ENEA MAFFEI • Focal Area Infection Biology, Biozentrum, University of Basel, Basel, Switzerland PABLO MANFREDI • Biozentrum, University of Basel, Basel, Switzerland CHARLOTTE MICHAUX • Section of Microbiology, Medical Research Council Centre for Molecular Bacteriology and Infection, Imperial College London, London, UK; Department of Microbiology, Harvard Medical School, Boston, MA, USA SOFIYA MICHEVA-VITEVA • Bioscience Division, Los Alamos National Laboratory, Los Alamos, NM, USA JAN MICHIELS • VIB-KU Leuven Center for Microbiology, VIB, Leuven, Belgium; Centre of Microbial and Plant Genetics, KU Leuven, Leuven, Belgium SAYED GOLAM MOHIUDDIN • Department of Chemical and Biomolecular Engineering, University of Houston, Houston, TX, USA STEFANY MORENO-GA´MEZ • Department of Environmental Systems Science, Institute of Biogeochemistry and Pollutant Dynamics, ETH Zurich, Zurich, Switzerland; Department of Environmental Microbiology, Eawag, Du¨bendorf, Switzerland; Groningen Institute for Evolutionary Life Sciences, University of Groningen, Groningen, The Netherlands ALLISON M. MURAWSKI • Department of Molecular Biology, Princeton University, Princeton, NJ, USA; Rutgers Robert Wood Johnson Medical School, Piscataway, NJ, USA JOSHUA P. M. NEWSON • Department of Biology, Institute of Microbiology, ETH Zurich, Zu¨rich, Switzerland TIEP K. NGUYEN • Pharmacologie Cellulaire et Mole´culaire, Louvain Drug Research Institute, Universite´ catholique de Louvain, Brussels, Belgium MEHMET A. ORMAN • Department of Chemical and Biomolecular Engineering, University of Houston, Houston, TX, USA FRE´DE´RIC PEYRUSSON • Pharmacologie Cellulaire et Mole´culaire, Louvain Drug Research Institute, Universite´ catholique de Louvain, Brussels, Belgium LAUREN C. RADLINSKI • Department of Microbiology and Immunology, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA JAKUB L. RADZIKOWSKI • Molecular Systems Biology, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, Groningen, The Netherlands; Centre for Engagement and Simulation Science (ICCESS), Department of Surgery and Cancer, Faculty of Medicine, Imperial College London, London, UK KATHERINE RITTENBACH • Department of Molecular Biology, Princeton University, Princeton, NJ, USA SE´VERIN RONNEAU • Section of Microbiology, Medical Research Council Centre for Molecular Bacteriology and Infection, Imperial College London, London, UK; Department of Microbiology, Harvard Medical School, Boston, MA, USA SARAH E. ROWE • Department of Microbiology and Immunology, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA ISABELLA SANTI • Biozentrum, University of Basel, Basel, Switzerland

Contributors

xi

ALEXANDER SCHMIDT • Biozentrum, University of Basel, Basel, Switzerland CRISTINA SERAL • Pharmacologie Cellulaire et Mole´culaire, Louvain Drug Research Institute, Universite´ catholique de Louvain, Brussels, Belgium; Department of Microbiology, Hospital Clı´nico Universitario Lozano Blesa, Zaragoza, Spain ASHELYN E. SIDDERS • Department of Microbiology and Immunology, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA EMRAH S¸IMS¸EK • Department of Physics, Emory University, Atlanta, GA, USA; Department of Biomedical Engineering, Duke University, Durham, NC, USA PAUL M. TULKENS • Pharmacologie Cellulaire et Mole´culaire, Louvain Drug Research Institute, Universite´ catholique de Louvain, Brussels, Belgium FRANC¸OISE VAN BAMBEKE • Pharmacologie Cellulaire et Mole´culaire, Louvain Drug Research Institute, Universite´ catholique de Louvain, Brussels, Belgium LAURENCE VAN MELDEREN • Cellular and Molecular Microbiology, Faculte´ des Sciences, Universite´ Libre de Bruxelles (ULB), Gosselies, Belgium SIMON VAN VLIET • Biozentrum, University Basel, Basel, Switzerland BRAM VAN DEN BERGH • VIB-KU Leuven Center for Microbiology, KU Leuven Centre of Microbial and Plant Genetics, Leuven, Belgium SILKE R. VEDELAAR • Molecular Systems Biology, Groningen Biomolecular Sciences and Biotechnology Institute, University of Groningen, Groningen, The Netherlands LAURE VERSTRAETE • VIB-KU Leuven Center for Microbiology, VIB, Leuven, Belgium; Centre of Microbial and Plant Genetics, KU Leuven, Leuven, Belgium NATALIE VERSTRAETEN • VIB-KU Leuven Center for Microbiology, VIB, Leuven, Belgium; Centre of Microbial and Plant Genetics, KU Leuven, Leuven, Belgium GANG WANG • Pharmacologie Cellulaire et Mole´culaire, Louvain Drug Research Institute, Universite´ catholique de Louvain, Brussels, Belgium DORIEN WILMAERTS • VIB-KU Leuven Center for Microbiology, KU Leuven Centre of Microbial and Plant Genetics, Leuven, Belgium ETTHEL WINDELS • VIB-KU Leuven Center for Microbiology, KU Leuven Centre of Microbial and Plant Genetics, Leuven, Belgium

Part I Introduction

Chapter 1 Studying Bacterial Persistence: Established Methods and Current Advances Elen Louwagie , Laure Verstraete , Jan Michiels , and Natalie Verstraeten Abstract To date, we are living in a postantibiotic era in which several human pathogens have developed multidrug resistance and very few new antibiotics are being discovered. In addition to the problem of antibiotic resistance, every bacterial population harbors a small fraction of transiently antibiotic-tolerant persister cells that can survive lethal antibiotic attack. Upon cessation of the treatment, these persister cells wake up and give rise to a new, susceptible population. Studies conducted over the past two decades have demonstrated that persister cells are key players in the recalcitrance of chronic infections and that they contribute to antibiotic resistance development. As a consequence, the scientific interest in persistence has increased tremendously and while some questions remain unanswered, many important insights have been brought to light thanks to the development of dedicated techniques. In this chapter, we provide an overview of wellestablished methods in the field and recent advances that have facilitated the investigation of persister cells and we highlight the challenges to be tackled in future persistence research. Key words Bacterial persistence, Persistence, Antibiotic tolerance, Antibiotics

1

Bacterial Persistence In 1942, while studying the mode of action of the just recently discovered penicillin, Hobby and coworkers noted that the killing rate of streptococci decreases over time and that 1% of the bacterial population survives even extended treatment [1]. Two years later, Bigger observed that also a small fraction of staphylococci bacteria withstand penicillin treatment and he called the surviving cells “persisters” [2]. He stated that persisters are in a dormant and nondividing state which is not heritable to the progeny, clearly distinguishing these cells from resistant mutants. While Bigger

*Authors Elen Louwagie and Laure Verstraete and Authors Jan Michiels and Natalie Verstraeten contributed equally to this work. Natalie Verstraeten and Jan Michiels (eds.), Bacterial Persistence: Methods and Protocols, Methods in Molecular Biology, vol. 2357, https://doi.org/10.1007/978-1-0716-1621-5_1, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2021

3

4

Elen Louwagie et al.

already pointed out that persistence could contribute to the failure of antibiotic treatment, the scientific interest in persister cells remained virtually nonexistent for 40 years following their discovery [2]. Moyed and Bertrand revived attention to persistence research in 1983 with the introduction of the so-called high persistence (hip) mutants [3]. These mutants were originally obtained by chemical mutagenesis and some of them displayed a 10,000-fold increase in persister fraction compared to the parental Escherichia coli K-12 strain. The observation of the high-persister phenotype was not restricted to β-lactam antibiotics, indicating for the first time that persistence is not specific to one antibiotic [3]. This incentive, combined with the increasing awareness of the clinical implications of persistence, encouraged other research groups to study persistence, leading to several groundbreaking results over the past years which are summarized in some excellent reviews [4– 13]. To date, multiple approaches and experimental setups have been adopted to understand the characteristics of persister cells, resulting in different perspectives on the matter. In 2018, the first international workshop on “Bacterial Persistence and Antimicrobial Therapy” was held in Ascona, Switzerland, and this has led to the publication of a general definition of persistence and guidelines for how to study it [10]. Antibiotic persistence is generally described as the ability of a bacterial population to survive exposure to bactericidal antibiotics. In contrast to population-wide tolerance, this ability is restricted to a small subpopulation of cells within an isogenic population. The killing rate of this isogenic population is thus biphasic with the antibiotic-sensitive subpopulation rapidly being killed, followed by the killing of the antibiotic-tolerant persister subpopulation at a much slower pace (Fig. 1) [10]. Persistence must be distinguished from resistance which is the survival and growth of cells during antibiotic stress due to the acquisition of stable genetic mutations. Persistence is also different from heteroresistance which refers to a subpopulation of cells that survive antibiotic treatment due to their transiently resistant phenotype. Noteworthy, persister cells cannot replicate in the presence of antibiotics while heteroresistant cells can. However, as is the case with heteroresistant cells, the drug insusceptibility of persisters is reversible [14]. Indeed, persister cells can switch back to the antibiotic-sensitive state and resume growth upon antibiotic removal, yielding a new, susceptible population. Persister cells are stochastically generated or formed as a response to environmental changes such as antibiotic or nutrient stress. The exact mechanisms of persister formation are still not clear, although dormancy or inactivity of the antibiotic target have often been linked with persistence. The presence of persister cells has been detected both in vitro and in vivo in several bacterial species [10, 11, 15]. Moreover, a similar phenomenon has been observed

Bacterial Persistence: Methods and Advances

5

Fig. 1 Survival kinetics of an isogenic population during antibiotic treatment. The bulk population is rapidly killed over time, followed by a slower killing rate due to the presence of antibiotic-tolerant persister cells. Upon removal of the antibiotic and regrowth in fresh medium, persister cells revert to the normal growing state, generating a population that is equally sensitive to antibiotics as the original population

in eukaryotes, which indicates that persistence is a general survival strategy [16, 17].

2

Clinical Relevance In the past few decades, persister cells have been recognized as key players in the recalcitrance of chronic infections, thereby putting a significant burden on human health [18]. Most of these infections are biofilm-related. For example, Pseudomonas aeruginosa, Candida albicans, uropathogenic E. coli, Salmonella species, and Staphylococcus aureus are able to attach to the surfaces of organ tissues and/or indwelling devices where they form a biofilm [19–23], that is, a structured microbial community embedded in a self-produced matrix consisting of polysaccharides, lipids, proteins, and DNA [24]. Additionally, uropathogenic E. coli is able to invade underlying bladder epithelial cells where it forms quiescent intracellular reservoirs with biofilm-like properties [25]. Within a biofilm environment, pathogens are shielded from the immune system, leaving clearance of the infection to be mostly dependent on antimicrobial treatment. However, effective clearing of the infection is often not achievable with these treatments. For a long time, it was thought that treatment failure was due to impaired penetrance of antimicrobials in the biofilm, but evidence is mounting that biofilms also

6

Elen Louwagie et al.

harbor an increased number of persister cells that are tolerant to the applied antimicrobials [16, 26–30]. A first correlation was shown ten years ago through the identification of hip mutants amongst clinical isolates originating from cystic fibrosis patients with chronic P. aeruginosa infections [31] and patients with chronic C. albicans infections [32]. Recently, similar evidence was provided for recurrent urinary tract infections caused by uropathogenic E. coli [33]. Persister cells have also been shown to be important in non–biofilm-related chronic infections. Well-known examples are Mycobacterium tuberculosis infections, in which M. tuberculosis invades and reprograms alveolar macrophages and resides in a tolerant, dormant state inside granulomas, while being protected from immune responses [34], a strategy that is also used by Salmonella species [35]. Recently, a screen of M. tuberculosis clinical isolates revealed the presence of hip mutants [36]. Besides being responsible for the recalcitrant nature of important chronic infections, persistence has been shown to pave the way for resistance development, which is recognized as one of the biggest threats to human health [37]. Both the prolonged presence of the infecting pathogen under intermittent antibiotic treatment and the increased mutation rates associated with the persister phenotype enhance the development of resistance-conferring mutations [38–40]. Finally, parallels have been drawn between microbial persistence and the presence of tolerant subpopulations in cancer tumors [41– 45]. Altogether, persistence plays a pivotal role, not only in chronic infections, but also in resistance development and the outcome of cancer treatments, which demonstrates the importance of persistence research.

3

Challenges in Persistence Research Although recent years have seen a tremendous increase in persistence research, the interest in the phenomenon is of a relatively new nature. Not only has the clinical relevance of persistence remained highly underestimated for many years, the study of persister cells has also been impeded by the many difficulties that are inherently linked to the phenotype. The continuous switching between the normal and persister state gives rise to a variable yet usually very small persister fraction, typically within a range of 0.001% to 1% of the entire population [9]. Moreover, this subpopulation is characterized by heterogeneity in cell physiologies and tolerance to antibiotics [46]. These issues have been tackled by major advances in microscopic and omics technologies, and the enrichment and isolation of persisters. New microscopic techniques allow researchers to study persistence at the single-cell level using fluorescent (timelapse) microscopy, microfluidics, and flow cytometry, which yields valuable information on cell growth and cell death, metabolic state and gene expression of individual cells [14]. In these single-cell

Bacterial Persistence: Methods and Advances

7

studies, as well as in proteome and transcriptome analyses, the enrichment and isolation of persisters is crucial given the low number of persister cells. One commonly used approach is to lyse all antibiotic-sensitive cells by antibiotic treatment and harvest the nonlysed, antibiotic-tolerant persister cells by centrifugation [47, 48] or the use of magnetic beads [49]. These methods rely on the killing kinetics of antibiotics and thus on the physiological state of the bacteria which is circumvented by lysing normal cells using a mixture of alkaline and enzymatic solutions [50]. Another approach is based on the assumption that persister cells are in a dormant, nongrowing state and/or that growth-related genes are not, or to a lesser extent, expressed in persister cells. By the use of fluorescent reporters such as the green fluorescent protein (GFP), persister cells can be sorted from the bulk population by fluorescence activated cell sorting (FACS) [51] or quantified by flow cytometry [52–55]. New isolation and enrichment protocols are based on the filamentation of β-lactam-treated bacteria followed by filtration to isolate persister cells [56] or cyclic exposure to antibiotics [57]. Prior to isolation, the formation of persister-like cells can be induced by antibiotic or chemical pretreatment [58, 59] or toxin overexpression [60]. Since the persister population is heterogeneous and variable, and most enrichment and isolation methods rely on persister physiology, persistence should be studied by a combination of population and single-cell approaches which both will benefit from future technological advances. Currently used methods and recent advances to study persistence are briefly introduced below.

4

Current Methods

4.1 Genetic Screenings

One of the first methods to be developed for the identification of new persister mechanisms was the large-scale screening of transposon, knockout, and overexpression libraries. In general, strains from a mutant library are exposed to antibiotics for several hours and the number of surviving cells is determined. Strains with an increased or decreased survival presumably contain one or multiple mutated or overexpressed genes involved in persistence and are further tested for sensitivity to other antibiotics and antibiotic resistance. Two research groups have independently screened E. coli mini-Tn10 insertion libraries, both times yielding a few knockout mutants with reduced persister levels upon antibiotic exposure [61, 62]. Also in other organisms such as M. tuberculosis, P. aeruginosa, and S. aureus, transposon mutagenesis has been performed to gain new insights in general persistence pathways [63– 66]. The E. coli K-12 knockout library (Keio collection) [67] has also been subjected to persistence screening by two independent groups using different antibiotics and growth conditions, leading to

8

Elen Louwagie et al.

the identification of several novel persister genes [68, 69]. The abovementioned screenings allowed to identify only nonessential genes and their potential role in persistence. To discover genes that are indispensable for growth as well as regulatory RNAs with a role in persistence, screening of an overexpression library has proven instrumental [68, 70]. Overexpression of two proteins involved in glycerol metabolism resulted in an increase of persister survival in E. coli [71]. Other gain-of-function screenings also showed a link between energy production and increased persistence [72, 73]. Finally, due to recent advances in next-generation sequencing, TnSeq [74] and Barseq [17] have emerged as popular high-throughput screening methods in the search for new persister genes. While genetic screenings have led to the discovery of many new persister genes, the role of these genes in persister formation and/or maintenance is in many cases still not clear. Importantly, the genome-wide screenings revealed mostly global regulators and genes involved in cell metabolism and did not produce any mutants unable to form persisters. This strongly suggests that there is not just one single mechanism underlying persistence and that several molecular pathways are involved. 4.2 Experimental Evolution

Several studies have shown that persistence can increase the fitness of the entire bacterial population in stressful conditions [75–77] and could thus be a phenotypic trait that can be selected for. Indeed, evolution experiments have shown that the number of persister cells can be tuned by the frequency of antibiotic exposure. In these experiments, bacterial cultures are intermittently exposed to high-dose antibiotics and regrown upon suspension in fresh medium, reflecting the clinical settings in which patients are regularly treated with antibiotics. It is therefore assumed that the studied persister cells and associated mutations are similar to those encountered in natural environments. Experimental evolution was already performed by Moyed and Bertrand in 1983 when they periodically treated E. coli populations with ampicillin, resulting in the evolution of hip mutants including the hipA7 strain that has since been widely used in persister research (see e.g. Subheadings 4.3.1 and 4.4.1 below) [3]. Later, two important evolution experiments were performed showing a rapid evolution toward increased persister levels [78] and tolerance [79] in E. coli without antibiotic resistance development. Different conditions (e.g., growth phase of the population prior to treatment, antibiotics, treatment duration) were used in these studies, but both experiments showed that in some of the evolved clones only a single point mutation was responsible for the increase in survival [78, 79]. Noteworthy, elevated persister levels in the notorious ESKAPE pathogens [80] and evolution to tolerance in S. aureus [81] upon periodic application of antibiotics were also observed. Some recent evolution experiments were also combined with proteomics or transcriptomics

Bacterial Persistence: Methods and Advances

9

analyses, allowing to study the proteome and transcriptome of the evolved mutants [57, 82]. Finally, experimental evolution has been used to study the link between persistence and resistance [39, 83]. 4.3 Transcriptomics and Proteomics

In addition to genetic screenings and experimental evolution, transcriptomics and proteomics have found their way to persistence research. This is not surprising, as persister cells are phenotypic variants within an isogenic population and thus differ from the majority of the population in transcript and protein contents. However, as the persister population comprises a small fraction of the total population, their transcriptome and proteome profiles can only be obtained after increasing the persister fraction and/or after proper enrichment or isolation of persister cells.

4.3.1 Transcriptomics

In 2004, the Lewis group was the first to study the expression profile of persister cells [47]. To cope with the low abundance of persisters, experiments were conducted with an E. coli strain carrying the hipA7 allele. To enrich the persister cells for transcriptome analysis, an exponential-phase culture was treated with the β-lactam antibiotic ampicillin and nonlysed persister cells were isolated through centrifugation. RNA was extracted from the isolated persister cells, enriched for mRNA, converted to cDNA, hybridized to a microarray and analyzed. The expression profile of persister cells showed a downregulation of metabolism and flagellar synthesis pathways and an upregulation of toxin–antitoxin (TA) modules and other proteins that slow down active bacterial processes such as translation and replication, all pointing toward a dormant state [47]. However, it should be noted that this isolation method has its downsides, as the antibiotic treatment prior to isolation could affect the expression profile. A few years later, the same research group exploited its findings to develop an alternative method for persister isolation from a wild-type E. coli population, without involvement of antibiotic treatment, that is based on the assumption of persister cells being dormant and thus having their translation machinery shut down [51]. Briefly, a degradable GFP was placed under the control of a ribosomal promoter and dim, growth-arrested cells were isolated using FACS. This isolation method proved its effectiveness, as the dim population was shown to be tolerant to antibiotics. Next, the expression profile of the dim persister population was determined as described before, which confirmed their previous findings [47, 51]. In 2011, the Lewis group also succeeded in determining the expression profile of M. tuberculosis exponentialphase persisters using similar techniques [84]. The method used to isolate persisters was again based on lysis of nonpersister cells but now using the cell wall-acting antibiotic D-cycloserine, followed by centrifugation. This study showed the expression profile of M. tuberculosis persisters to be very similar to the one of E. coli persisters [84]. Microarray hybridization was also used by the

10

Elen Louwagie et al.

Michiels group to investigate the role of DnpA, a de-N-acetylase, in P. aeruginosa persister formation by comparing expression profiles of the wild-type strain, a dnpA mutant and a dnpA overexpression strain, of which the latter causes an increased persister fraction. From this, the authors learnt that dnpA overexpression coincides with decreased amino acid biosynthesis and metabolism, which is in accordance with the studies mentioned above [85]. Later, transcriptome profiling of persisters was also done in other microorganisms, but then using the state-of-the-art RNA-sequencing technique instead of microarray hybridization [86–88]. Here, a library of cDNA fragments with ligated adapters is generated and subsequently sequenced using one of the currently existing highthroughput sequencing platforms, such as Illumina [89]. In a study with Staphylococcus epidermidis, RNA-seq analysis of biofilms containing different portions of dormant bacteria revealed that mostly genes related to oxidation-reduction processes and acetyl-CoA metabolic reactions are differently expressed in persister cells [86]. The same technology was also used to compare the expression profiles of Listeria monocytogenes nisin-treated and untreated cultures, showing persister cells to have a decreased energy production and flagellar synthesis [87]. Finally, RNA-seq analysis of macrophages infected with Salmonella Typhimurium revealed that residing persister cells have indeed ceased growth, but instead of being completely dormant, they still show transcriptional and translational activity enabling them to reprogram their macrophage host [88]. 4.3.2 Proteomics

Proteome profiling of persister cells holds great promise to gain insight in the persister phenotype because it allows for protein identification and quantification as well as the identification of interaction partners and post-translational modifications, which could reveal hidden levels of regulation. However, as standard proteomic analyses require high cell numbers and the number of persister cells in a culture is usually low, these types of studies are especially challenging. This explains why the number of reports applying proteomics to persister populations is rather limited. Moreover, the few studies that have been conducted have focused almost exclusively on M. tuberculosis and C. albicans persister cells due to their abundant presence inside granulomas and biofilms, respectively, and their ease of isolation [90]. M. tuberculosis resides in a nonreplicating persistent state inside granulomas, an environment that is characterized by hypoxia and nutrient starvation [91]. By imposing 6 weeks of nutrient starvation, these conditions were mimicked in vitro and then the proteome of both log-phase and nutrient-starved cells was analyzed through two-dimensional gel electrophoresis combined with liquid chromatography tandem mass spectrometry (LC-MS/MS) and/or matrix-assisted laser

Bacterial Persistence: Methods and Advances

11

desorption/ionization MS (MALDI-MS). These analyses revealed an increased abundance of TA modules, lipoproteins, and stringent response effectors and a decreased abundance of proteins involved in energy metabolism and amino acid, protein, and lipid biosynthesis [92, 93], thereby complementing the results obtained through transcriptomics studies [84]. In 2015, two research groups induced the nonreplicating persistent state by applying hypoxia and subsequently analyzed the proteome profile of these M. tuberculosis cells, as well as of reoxygenated, reactivated cells and exponential-phase cells grown in normoxic conditions, through LC-MS/MS or SWATH-MS [94, 95], thereby confirming previous findings [92, 93]. In addition to M. tuberculosis granulomas, C. albicans biofilms have also been shown to harbor an increased number of persister cells [16, 96–98], although it should be noted that two studies are now contradicting this statement [99, 100]. Similar to studies conducted in M. tuberculosis, LC-MS/MS was used to investigate the proteome of persister-enriched C. albicans biofilm populations, revealing an overall downregulation of metabolic processes, an upregulation of some genes related to the glyoxylate pathway, growth, virulence and the stress response [96], and a crucial role for alkyl hydroperoxide reductase 1 (AHP1) in persistence against amphotericin B [97]. Although challenging, methods have recently been developed to determine the proteomic profile of persisters in other microorganisms, such as E. coli. As their abundance is typically low, the number of persister cells should be increased or persister cells should be isolated prior to proteome analysis. As such, the proteome of nutrient-shift-induced E. coli persister cells could be investigated [101]. In another study, the persister fraction of E. coli populations was increased through evolution under cyclic antibiotic treatment and then proteomic analysis was performed on the evolved stationary-phase populations [57]. 4.4 Single-cell Techniques

Given the low abundance of persister cells and the transient and heterogeneous nature of the persister phenotype, persistence research greatly benefits from existing single-cell techniques such as single-cell time-lapse microscopy and flow cytometry.

4.4.1 Single-cell Time-Lapse Microscopy

Single-cell time-lapse microscopy allows to monitor the behavior of single cells over time through microscopic image analysis. In a pioneering study, Balaban [102] were the first to image single cells of an E. coli bacterial culture in a microfluidics device through the phases of treatment, in which the sensitive cells of the population were lysed through the action of ampicillin, followed by persister awakening and outgrowth in fresh medium. To avoid cells overgrowing each other and thus loosing track of cells, the device was patterned with narrow grooves in which the cells were loaded [102]. To cope with the issue of low abundance of persister cells,

12

Elen Louwagie et al.

the experiments were conducted with a hipA7 mutant. This system allowed the authors to build a mathematical model describing the switching rates between sensitive and persister cells [102]. Afterward, microfluidic devices have also been used to follow fluorescent protein induction in both sensitive and persister cells upon addition of fresh medium, before and after treatment, to further characterize the dormant persister state [103], to study the role of indole in persister formation [104], and to study the awakening mechanism of HokB-induced persister cells [105]. Despite the different strategies that are being used to increase the number of persister cells and the persister isolation methods that have been developed, a major downside of single-cell time-lapse microscopy of persister cells remains its relatively low throughput. Moreover, the existing microfluidic configurations do not allow to select and sort out specific single cells for further investigation. 4.4.2 Flow Cytometry

These throughput and sorting issues are circumvented by flow cytometry, a method that allows to assay thousands or even millions of single cells based on fluorescent measurements, followed by FACS to sort out the cells of interest. Although flow cytometry does not allow to track single cells over time, it has been proven very useful to study the persister phenotype. First of all, flow cytometry can be used to show the presence of a persister population within a culture. For example, E. coli stationary-phase populations were investigated with flow cytometry using a GFP dilution method and the presence of a bright, nongrowing persister population was confirmed [52]. Similarly, dual DsRed-GFP dilution and TIMERbac were used to study Salmonella subpopulations [53, 54]. Importantly, not only can persister populations be identified, but they can also be discerned from the viable but nonculturable (VBNC) cells within a population [55]. Second, persister awakening kinetics can be deduced from flow cytometry data using these fluorescent protein dilution methods [55, 106, 107]. Third, flow cytometry combined with FACS has become the state-of-the-art technique to study persister physiology. For example, by using the metabolic dye redox sensor green, it was shown that dormancy is not sufficient nor required for a cell to become a persister cell [108]. Alternatively, persister physiology can be studied by using fluorescent reporter fusions and screening the sorted populations for persistence [51, 109, 110]. Lastly, as already mentioned earlier, FACS can be used for isolation of persister cells [51].

4.4.3 Advances in Single-cell Technologies

To date, there has only been one study combining highthroughput and the ability to track and recover cells of interest. In this study, a carbenicillin-treated P. aeruginosa culture resuspended and diluted in fresh medium was applied to a hydrophilicin-hydrophobic patterned array, resulting in 106 dome-shaped

Bacterial Persistence: Methods and Advances

13

femtoliter droplets, each containing maximally one persister cell. After overnight incubation of the array, the cells that had divided were picked up using an automated micropipette and were checked for antibiotic susceptibility to exclude antibiotic resistance development [111, 112]. It is the rise of such innovative technologies that will provide great potential for future persistence research. 4.5 Mathematical Modelling

The outcome of wet lab experiments greatly depends on the used experimental procedures, often resulting in discrepancies in reported results when different methods are used. This is especially true for persistence as it is a complex phenotype to detect, quantify and study in detail [10, 113]. A major advantage of modelling is that wet lab experiments are strictly not necessary but rather used to support experimental outcomes or test new hypotheses. Moreover, predictive models can take into account clinically relevant antibiotic treatment profiles [40, 114] and the effect of the immune system and other host factors [115]. Two major types of models have been implemented in persistence research: population models, which describe the dynamics of populations and subpopulations, and mechanistic models, which focus on molecular mechanisms at the single-cell level. One of the first population models by Balaban [102] described the stochastic switching between normal and persister cells assuming constant rates of persister cell formation and awakening [102]. Later studies have used and extended this basic model to predict persister levels in different fluctuating conditions [75, 116, 117]. Interestingly, these models show that persister levels change according to the frequency of the antibiotic treatment [75, 78, 116–118], which is in line with empirical observations [78, 80]. An alternative population model proposes the asymmetric ageing of bacteria as a mechanism to explain the presence of persister cells [119]. More recent research focuses on the prediction of antibiotic resistance development from a persister population [39, 40, 114, 120]. Mechanistic models, on the other hand, describe the mechanisms responsible for the coexistence of normal, antibiotic-sensitive cells and antibiotic-tolerant cells within an isogenic population. These models explain phenotypic bistability by the stochastic expression of genes linked with persister formation and resuscitation, which is amplified by feedback loops. For example, the differential expression of toxin–antitoxin modules has been extensively studied in the past (e.g., [121–124]). However, the complete regulatory network is not unraveled yet and the identification of novel persister genes is essential to explain the observed phenotypic variability by mechanistic models. Even though predictive models provide us with crucial insights in persistence, most models focus on stochastic switching of persisters while the switching in response to environmental stimuli is often not considered. This leaves a gap in the current mathematical

14

Elen Louwagie et al.

predictions. Moreover, additional in vitro and in vivo observations are required as they could improve existing models [125]. 4.6

5

In vivo Models

A key question in persistence research is whether in vitro and in silico findings can be extrapolated to in vivo settings. In vivo models are not only interesting to study persister mechanisms but could also be a direct proof of the role of persister cells in relapsing infections. Several research groups have developed intracellular, biofilm, and chronic infection models to study persistence and most of the models are discussed in the review of Van den Bergh et al (2017) [9]. In some of these models, persister determinants identified in vitro could be validated. For example, a P. aeruginosa ΔrelA spoT mutant displays lower persister levels compared to the wild type upon treatment with ofloxacin in laboratory conditions. Similarly, infection by this mutant in murine intraperitoneal and subcutaneous biofilm infection models results in increased ofloxacin activity as exemplified by the increased mice survival and lower CFU counts, respectively [126]. However, some of the animal studies show discrepancies between in vitro and in vivo results. A Salmonella Δlon mutant, for example, does not alter the persistent phenotype in vitro, while infection by the same mutant in an intracellular macrophage model results in reduced persister formation compared to the wild-type strain [35]. Similarly, mutations that increase and decrease persistence in vivo were identified by infecting mice with a M. tuberculosis transposon library, followed by isoniazid treatment. These mutations, however, did not alter in vitro persister levels [127]. These examples show that in vivo models are crucial to study complex and condition-dependent phenotypes such as persistence. In addition to elucidating the characteristics and clinical relevance of persistence, in vivo models have been used to test the efficacy of new antipersister therapies [127–135], although animal models to specifically evaluate potential compounds do not exist yet.

Future Perspectives The methods described above have all contributed to the current knowledge of persistence. However, all of these techniques have their advantages and disadvantages and many questions remain unanswered. Given the challenges inherent to persistence research, that is, the heterogeneous and stochastic nature of the persister phenotype, and the shortcomings of the current persister isolation methods, there is a clear need to explore and implement highthroughput, single-cell techniques and to optimize representative in vivo models. Future advances in persistence methodologies will lead to new insights, thereby enhancing the development of antipersister therapies. The latter will not only allow effective treatment

Bacterial Persistence: Methods and Advances

15

of otherwise chronic infections, but will also slow down resistance development which will be crucial in the postantibiotic era we live in.

Acknowledgements E.L. and L.V. are Research Foundation Flanders (FWO)-fellows. The work was supported by grants from the FWO (G055517N, G0B0420N), KU Leuven (C16/17/006), and the Flanders Institute for Biotechnology (VIB). References 1. Hobby GL, Meyer K, Chaffee E (1942) Observations on the mechanism of action of penicillin. Exp Biol Med 50:281–285 2. Bigger JW (1944) Treatment of staphylococcal infections with penicillin. Lancet 244:497–500 3. Moyed HS, Bertrand KP (1983) hipA, a newly recognized gene of Escherichia coli K-12 that affects frequency of persistence after inhibition of murein synthesis. J Bacteriol 155:768–775 4. Lewis K (2010) Persister cells. Annu Rev Microbiol 64:357–372 5. Amato SM, Fazen CH, Henry TC et al (2014) The role of metabolism in bacterial persistence. Front Microbiol 5:70 6. Harms A, Maisonneuve E, Gerdes K (2016) Mechanisms of bacterial persistence during stress and antibiotic exposure. Science 354: aaf4268 7. Brauner A, Fridman O, Gefen O, Balaban NQ (2016) Distinguishing between resistance, tolerance and persistence to antibiotic treatment. Nat Rev Microbiol 14:320–330 8. Radzikowski JL, Schramke H, Heinemann M (2017) Bacterial persistence from a systemlevel perspective. Curr Opin Biotechnol 46:98–105 9. Van den Bergh B, Fauvart M, Michiels J (2017) Formation, physiology, ecology, evolution and clinical importance of bacterial persisters. FEMS Microbiol Rev 41:219–251 10. Balaban NQ, Helaine S, Lewis K et al (2019) Definitions and guidelines for research on antibiotic persistence. Nat Rev Microbiol 17:441–448 11. Wilmaerts D, Windels EM, Verstraeten N, Michiels J (2019) General mechanisms leading to persister formation and awakening. Trends Genet 35:401–411

12. Gollan B, Grabe G, Michaux C, Helaine S (2019) Bacterial persisters and infection: past, present, and progressing. Annu Rev Microbiol 73:359–385 13. Bakkeren E, Diard M, Hardt W-D (2020) Evolutionary causes and consequences of bacterial antibiotic persistence. Nat Rev Microbiol 18:479–490 14. Dewachter L, Fauvart M, Michiels J (2019) Bacterial heterogeneity and antibiotic survival: understanding and combatting persistence and heteroresistance. Mol Cell 76:255–267 15. Lewis K (2007) Persister cells, dormancy and infectious disease. Nat Rev Microbiol 5:48–56 16. LaFleur MD, Kumamoto CA, Lewis K (2006) Candida albicans biofilms produce antifungal-tolerant persister cells. Antimicrob Agents Chemother 50:3839–3846 17. Bojsen R, Regenberg B, Gresham D, Folkesson A (2016) A common mechanism involving the TORC1 pathway can lead to amphotericin B-persistence in biofilm and planktonic Saccharomyces cerevisiae populations. Sci Rep 6:21874 18. Fauvart M, De Groote VN, Michiels J (2011) Role of persister cells in chronic infections: clinical relevance and perspectives on antipersister therapies. J Med Microbiol 60:699–709 19. Douglas LJ (2003) Candida biofilms and their role in infection. Trends Microbiol 11:30–36 20. Steenackers H, Hermans K, Vanderleyden J, De Keersmaecker SCJ (2012) Salmonella biofilms: an overview on occurrence, structure, regulation and eradication. Food Res Int 45:502–531

16

Elen Louwagie et al.

21. Mulcahy LR, Isabella VM, Lewis K (2014) Pseudomonas aeruginosa biofilms in disease. Microb Ecol 68:1–12 22. Speziale P, Pietrocola G, Foster TJ, Geoghegan JA (2014) Protein-based biofilm matrices in Staphylococci. Front Cell Infect Microbiol 4:171 23. Narayanan A, Nair MS, Muyyarikkandy MS, Amalaradjou MA (2018) Inhibition and inactivation of uropathogenic Escherichia coli biofilms on urinary catheters by sodium selenite. Int J Mol Sci 19:1703 24. Flemming H-C, Wingender J (2010) The biofilm matrix. Nat Rev Microbiol 8:623–633 25. Anderson GG, Martin SM, Hultgren SJ (2004) Host subversion by formation of intracellular bacterial communities in the urinary tract. Microbes Infect 6:1094–1101 26. Lewis K (2001) Riddle of biofilm resistance. Antimicrob Agents Chemother 45:999–1007 27. Lewis K (2005) Persister cells and the riddle of biofilm survival. Biochemist 70:267–274 28. Roberts ME, Stewart PS (2005) Modelling protection from antimicrobial agents in biofilms through the formation of persister cells. Microbiology 151:75–80 29. Spoering AL, Lewis K (2001) Biofilms and planktonic cells of Pseudomonas aeruginosa have similar resistance to killing by antimicrobials. J Bacteriol 183:6746–6751 30. Waters EM, Rowe SE, O’Gara JP, Conlon BP (2016) Convergence of Staphylococcus aureus persister and biofilm research: can biofilms be defined as communities of adherent persister cells? PLoS Pathog 12:e1006012 31. Mulcahy LR, Burns JL, Lory S, Lewis K (2010) Emergence of Pseudomonas aeruginosa strains producing high levels of persister cells in patients with cystic fibrosis. J Bacteriol 192:6191–6199 32. LaFleur MD, Qi Q, Lewis K (2010) Patients with long-term oral carriage harbor highpersister mutants of Candida albicans. Antimicrob Agents Chemother 54:39–44 33. Schumacher MA, Balani P, Min J et al (2015) HipBA-promoter structures reveal the basis of heritable multidrug tolerance. Nature 524:59–64 34. Gengenbacher M, Kaufmann SHE (2012) Mycobacterium tuberculosis: success through dormancy. FEMS Microbiol Rev 36:514–532 35. Helaine S, Cheverton AM, Watson KG et al (2014) Internalization of Salmonella by macrophages induces formation of nonreplicating persisters. Science 343:204–208

36. Torrey HL, Keren I, Via LE et al (2016) High persister mutants in Mycobacterium tuberculosis. PLoS One 11:e0155127 37. O’Neill J (2016) Tackling drug-resistant infections globally: final report and recommendations: the review on antimicrobial resistance. Available via https://www. biomerieuxconnection.com/wp-content/ uploads/2018/04/Tackling-Drug-Resis tant-Infections-Globally_-Final-Report-andRecommendations.pdf. Accessed 26 August 2020 38. Cohen NR, Lobritz MA, Collins JJ (2013) Microbial persistence and the road to drug resistance. Cell Host Microbe 13:632–642 39. Levin-Reisman I, Ronin I, Gefen O et al (2017) Antibiotic tolerance facilitates the evolution of resistance. Science 355:826–830 40. Windels EM, Michiels JE, Fauvart M et al (2019) Bacterial persistence promotes the evolution of antibiotic resistance by increasing survival and mutation rates. ISME 13:1239–1251 41. Sharma SV, Lee DY, Li B et al (2010) A chromatin-mediated reversible drug-tolerant state in cancer cell subpopulations. Cell 141:69–80 42. Chen J, Li Y, Yu TS et al (2012) A restricted cell population propagates glioblastoma growth after chemotherapy. Nature 488:522–526 43. Raha D, Wilson TR, Peng J et al (2014) The cancer stem cell marker aldehyde dehydrogenase is required to maintain a drug-tolerant tumor cell subpopulation. Cancer Res 74:3579–3590 44. Pribluda A, De La Cruz CC, Jackson EL (2015) Intratumoral heterogeneity: from diversity comes resistance. Clin Cancer Res 21:2916–2923 45. Terai H, Kitajima S, Potter DS et al (2018) ER stress signaling promotes the survival of cancer “persister cells” tolerant to EGFR tyrosine kinase inhibitors. Cancer Res 78:1044–1057 46. Allison KR, Brynildsen MP, Collins JJ (2011) Heterogeneous bacterial persisters and engineering approaches to eliminate them. Curr Opin Microbiol 14:593–598 47. Keren I, Shah D, Spoering A et al (2004) Specialized persister cells and the mechanism of multidrug tolerance in Escherichia coli. J Bacteriol 186:8172–8180 48. Pu Y, Zhao Z, Li Y et al (2016) Enhanced efflux activity facilitates drug tolerance in dormant bacterial cells. Mol Cell 62:284–294

Bacterial Persistence: Methods and Advances 49. Sulaiman JE, Hao C, Lam H (2018) Specific enrichment and proteomics analysis of Escherichia coli persisters from rifampin pretreatment. J Proteome Res 17:3984–3996 ˜as-Duarte SJ, Restrepo S, Pedraza JM 50. Can (2014) Novel protocol for persister cells isolation. PLoS One 9:e88660 51. Shah D, Zhang Z, Khodursky A et al (2006) Persisters: a distinct physiological state of E. coli. BMC Microbiol 6:53 ˜ ers A, Luidalepp H et al (2008) 52. Roostalu J, Jo Cell division in Escherichia coli cultures monitored at single cell resolution. BMC Microbiol 8:68 53. Helaine S, Thompson JA, Watson KG et al (2010) Dynamics of intracellular bacterial replication at the single cell level. Proc Natl Acad Sci U S A 107:3746–3751 54. Claudi B, Spro¨te P, Chirkova A et al (2014) Phenotypic variation of Salmonella in host tissues delays eradication by antimicrobial chemotherapy. Cell 158:722–733 55. Golam S, Kavousi P, Orman MA (2020) Flow-cytometry analysis reveals persister resuscitation characteristics. BMC Microbiol 20:202 56. Windels EM, Meriem Z Ben, Zahir T et al (2019) Enrichment of persisters enabled by a ß-lactam-induced filamentation method reveals their stochastic single-cell awakening. Commun Biol 2:426 57. Sulaiman JE, Lam H (2020) Proteomic investigation of tolerant Escherichia coli populations from cyclic antibiotic treatment. J Proteome Res 19:900–913 58. Do¨rr T, Lewis K, Vulic´ M (2009) SOS response induces persistence to fluoroquinolones in Escherichia coli. PLoS Genet 5: e1000760 59. Johnson PJT, Levin BR (2013) Pharmacodynamics, population dynamics, and the evolution of persistence in Staphylococcus aureus. PLoS Genet 9:e1003123 60. Mok WWK, Park JO, Rabinowitz JD, Brynildsen MP (2015) RNA futile cycling in model persisters derived from MazF accumulation. mBio 6:e01588-15 61. Li Y, Zhang Y (2007) PhoU is a persistence switch involved in persister formation and tolerance to multiple antibiotics and stresses in Escherichia coli. Antimicrob Agents Chemother 51:2092–2099 62. Hu Y, Coates ARM (2005) Transposon mutagenesis identifies genes which control antimicrobial drug tolerance in stationary-phase Escherichia coli. FEMS Microbiol Lett 243:117–124

17

63. Dhar N (2010) Mckinney JD. Mycobacterium tuberculosis persistence mutants identified by screening in isoniazid-treated mice 107:12275–12280 64. De Groote VN, Verstraeten N, Fauvart M et al (2009) Novel persistence genes in Pseudomonas aeruginosa identified by high-throughput screening. FEMS Microbiol Lett 297:73–79 65. Manuel J, Zhanel GG, De Kievit T (2010) Cadaverine suppresses persistence to carboxypenicillins in Pseudomonas aeruginosa PAO1. Antimicrob Agents Chemother 54:5173–5179 66. Wang W, Chen J, Chen G et al (2015) Transposon mutagenesis identifies novel genes associated with Staphylococcus aureus persister formation. Front Microbiol 6:1437 67. Baba T, Ara T, Hasegawa M et al (2006) Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol 2:2006.0008 68. Hansen S, Lewis K, Vulic´ M (2008) Role of gobal regulators and nucleotide metabolism in antibiotic tolerance in Escherichia coli. Antimicrob Agents Chemother 52:2718–2726 69. Ma C, Sim S, Shi W et al (2010) Energy production genes sucB and ubiF are involved in persister survival and tolerance to multiple antibiotics and stresses in Escherichia coli. FEMS Microbiol Lett 303:33–40 70. Lewis K (2008) Multidrug tolerance of biofilms and persister cells. In: Romeo T (ed) bacterial bioflms. Current topics in microbiology and immunology, vol 322. Springer, Heidelberg 71. Spoering AL, Vulic´ M, Lewis K (2006) GlpD and PlsB participate in persister cell formation in Escherichia coli. J Bacteriol 188:5136–5144 72. Leung V, Le´vesque CM (2012) A stressinducible quorum-sensing peptide mediates the formation of persister cells with noninherited multidrug tolerance. J Bacteriol 194:2265–2274 73. Girgis HS, Harris K, Tavazoie S (2012) Large mutational target size for rapid emergence of bacterial persistence. Proc Natl Acad Sci U S A 109:12740–12745 74. Shan Y, Lazinski D, Rowe S et al (2015) Genetic basis of persister tolerance to aminoglycosides in Escherichia coli. mBio 6:e0007815 75. Kussell E, Kishony R, Balaban NQ, Leibler S (2005) Bacterial persistence: a model of survival in changing environments. Genetics 169:1807–1814 76. Kussell E, Leibler S (2005) Phenotypic diversity, population growth, and information in

18

Elen Louwagie et al.

fluctuating environments. Science 309:2075–2078 77. Stewart B, Rozen DE (2011) Genetic variation for antibiotic persistence in Escherichia coli. Evolution 66:933–939 78. Van den Bergh B, Michiels JE, Wenseleers T et al (2016) Frequency of antibiotic application drives rapid evolutionary adaptation of Escherichia coli persistence. Nat Microbiol 1:16020 79. Fridman O, Goldberg A, Ronin I et al (2014) Optimization of lag time underlies antibiotic tolerance in evolved bacterial populations. Nature 513:418–421 80. Michiels JE, Van den Bergh B, Verstraeten N, Fauvart M (2016) In vitro emergence of high persistence upon periodic aminoglycoside challenge in the ESKAPE pathogens. Antimicrob Agents Chemother 60:4630–4637 81. Mechler L, Herbig A, Paprotka K et al (2015) A novel point mutation promotes growth phase-dependent daptomycin tolerance in Staphylococcus aureus. Antimicrob Agents Chemother 59:5366–5376 82. Khare A, Tavazoie S (2020) Extreme antibiotic persistence via heterogeneity-generating mutations targeting translation. mSystems 5: e00847-19 83. Windels EM, Michiels JE, Bergh B Van Den, Fauvart M (2019) Antibiotics: combatting tolerance to stop resistance. mBio 10: e02095–19 84. Keren I, Minami S, Rubin E, Lewis K (2011) Characterization and transcriptome analysis of Mycobacterium tuberculosis persisters. mBio 2: e00100-11 85. Liebens V, Defraine V, Ven der Leyden A et al (2014) A putative de-N-acetylase of the PIG-L superfamily affects fluoroquinolone tolerance in Pseudomonas aeruginosa. Pathog Dis 71:39–54 86. Carvalhais V, Franc¸a A, Cerca F et al (2014) Dormancy within Staphylococcus epidermidis biofilms: a transcriptomic analysis by RNA-seq. Appl Microbiol Biotechnol 98:2585–2596 87. Wu S, Yu P-L, Wheeler D, Flint S (2018) Transcriptomic study on persistence and survival of Listeria monocytogenes following lethal treatment with nisin. J Glob Antimicrob Resist 15:25–31 88. Stapels DAC, Hill PWS, Westermann AJ et al (2018) Salmonella persisters undermine host immune defenses during antibiotic treatment. Science 362:1156–1160 89. Wang Z, Gerstein M, Snyder M (2009) RNA-Seq: a revolutionary tool for transcriptomics. Nat Rev Genet 10:57–63

90. Sulaiman JE, Lam H (2019) Application of proteomics in studying bacterial persistence. Expert Rev Proteomics 16:227–239 91. Wayne LG (1994) Dormancy of Mycobacterium tuberculosis and latency of disease. Eur J Clin Microbiol Infect Dis 13:908–914 92. Betts JC, Lukey PT, Robb LC et al (2002) Evaluation of a nutrient starvation model of Mycobacterium tuberculosis persistence by gene and protein expression profiling. Mol Microbiol 43:717–731 93. Albrethsen J, Agner J, Piersma SR et al (2013) Proteomic profiling of Mycobacterium tuberculosis identifies nutrient-starvation-responsive toxin-antitoxin systems. Mol Cell Proteomics 12:1180–1191 94. Gopinath V, Raghunandanan S, Gomez RL et al (2015) Profiling the proteome of Mycobacterium tuberculosis during dormancy and reactivation. Mol Cell Proteomics 14:2160–2176 95. Schubert OT, Ludwig C, Kogadeeva M et al (2015) Absolute proteome composition and dynamics during dormancy and resuscitation of Mycobacterium tuberculosis. Cell Host Microbe 18:96–108 96. Li P, Seneviratne CJ, Alpi E et al (2015) Delicate metabolic control and coordinated stress response critically determine antifungal tolerance of Candida albicans biofilm persisters. Antimicrob Agents Chemother 59:6101–6112 97. Truong T, Zeng G, Lin Q et al (2016) Comparative ploidy proteomics of Candida albicans biofilms unraveled the role of the AHP1 gene in the biofilm persistence against amphotericin B. Mol Cell Proteomics 15:3488–3500 98. Zhigang JS, Haoyue L, Jing C et al (2016) Candida albicans amphotericin B-tolerant persister formation is closely related to surface adhesion. Mycopathologia 181:41–49 99. Al-Dhaheri RS, Douglas LJ (2008) Absence of amphotericin B-tolerant persister cells in biofilms of some Candida species. Antimicrob Agents Chemother 52:1884–1887 100. Denega I, D’Enfert C, Bachellier-Bassi S (2019) Candida albicans biofilms are generally devoid of persister cells. Antimicrob Agents Chemother 63:e01979–e01918 101. Radzikowski JL, Vedelaar S, Siegel D et al (2016) Bacterial persistence is an active σS stress response to metabolic flux limitation. Mol Syst Biol 12:882 102. Balaban NQ (2004) Bacterial persistence as a phenotypic switch. Science 305:1622–1625 103. Gefen O, Gabay C, Mumcuoglu M et al (2008) Single-cell protein induction

Bacterial Persistence: Methods and Advances dynamics reveals a period of vulnerability to antibiotics in persister bacteria. Proc Natl Acad Sci U S A 105:6145–6149 104. Vega NM, Allison KR, Khalil AS, Collins JJ (2012) Signaling-mediated bacterial persister formation. Nat Chem Biol 8:431–433 105. Wilmaerts D, Dewachter L, De Loose P-J et al (2019) HokB monomerization and membrane repolarization control persister awakening. Mol Cell 75:1031–1042 ˜ ers A, Kaldalu N, Tenson T (2010) The 106. Jo frequency of persisters in Escherichia coli reflects the kinetics of awakening from dormancy. J Bacteriol 192:3379–3384 ˜ ers A, Kaldalu N, Tenson T 107. Luidalepp H, Jo (2011) Age of inoculum strongly influences persister frequency and can mask effects of mutations implicated in altered persistence. J Bacteriol 193:3598–3605 108. Orman MA, Brynildsen MP (2013) Dormancy is not necessary or sufficient for bacterial persistence. Antimicrob Agents Chemother 57:3230–3239 109. Verstraeten N, Knapen WJ, Kint CI et al (2015) Obg and membrane depolarization are part of a microbial bet-hedging strategy that leads to antibiotic tolerance. Mol Cell 59:9–21 110. Henry TC, Brynildsen MP (2016) Development of Persister-FACSeq: a method to massively parallelize quantification of persister physiology and its heterogeneity. Sci Rep 6:25100 111. Sakakihara S, Araki S, Iino R, Noji H (2010) A single-molecule enzymatic assay in a directly accessible femtoliter droplet array. Lab Chip 10:3355–3362 112. Iino R, Matsumoto Y, Nishino K et al (2013) Design of a large-scale femtoliter droplet array for single-cell analysis of drug-tolerant and drug-resistant bacteria. Front Microbiol 4:300 113. Harms A, Fino C, Sørensen MA et al (2017) Prophages and growth dynamics confound experimental results with antibiotic-tolerant persister cells. mBio 8:e01964-17 114. Levin BR, Rozen DE (2006) Non-inherited antibiotic resistance. Nat Rev Microbiol 4:556–562 115. Ankomah P, Levin BR (2014) Exploring the collaboration between antibiotics and the immune response in the treatment of acute, self-limiting infections. Proc Natl Acad Sci U S A 111:8331–8338 116. Patra P, Klumpp S (2013) Population dynamics of bacterial persistence. PLoS One 8: e62814

19

117. Patra P, Klumpp S (2015) Emergence of phenotype switching through continuous and discontinuous evolutionary transitions. Phys Biol 12:046004 118. Gardner A, West SA, Griffin AS (2007) Is bacterial persistence a social trait? PLoS One 2:e752 119. Klapper I, Gilbert P, Ayati BP et al (2007) Senescence can explain microbial persistence. Microbiology 153:3623–3630 120. Levin-Reisman I, Brauner A, Ronin I, Balaban NQ (2019) Epistasis between antibiotic tolerance, persistence, and resistance mutations. Proc Natl Acad Sci 116:14734–14739 121. Feng J, Kessler DA, Ben-Jacob E, Levine H (2014) Growth feedback as a basis for persister bistability. Proc Natl Acad Sci U S A 111:544–549 122. Rotem E, Loinger A, Ronin I et al (2010) Regulation of phenotypic variability by a threshold-based mechanism underlies bacterial persistence. Proc Natl Acad Sci U S A 107:12541–12546 123. Lou C, Li Z, Ouyang Q (2008) A molecular model for persister in E. coli. J Theor Biol 255:205–209 124. Fasani RA, Savageau MA (2013) Molecular mechanisms of multiple toxin-antitoxin systems are coordinated to govern the persister phenotype Proc Natl Acad Sci U S A 110: E2528–E2537 ˜ ez-Guzma´n 125. Stepanyan K, Wenseleers T, Due´n EA et al (2015) Fitness trade-offs explain low levels of persister cells in the opportunistic pathogen Pseudomonas aeruginosa. Mol Ecol 24:1572–1583 126. Nguyen D, Joshi-Datar A, Lepine F et al (2011) Active starvation responses mediate antibiotic tolerance in biofilms and nutrientlimited bacteria. Science 334:982–986 127. Dhar N, Mckinney JD (2010) Mycobacterium tuberculosis persistence mutants identified by screening in isoniazid-treated mice. Proc Natl Acad Sci U S A 107:12275–12280 128. Chua SL, Kuok J, Yam H et al (2016) Selective labelling and eradication of antibiotictolerant bacterial populations in Pseudomonas aeruginosa biofilms. Nat Commun 7:10750 129. Briers Y, Walmagh M, Van Puyenbroeck V et al (2014) Engineered endolysin-based “Artilysins” to combat multidrug-resistant gram-negative pathogens. mBio 5:e0137914 130. Allison KR, Brynildsen MP, Collins JJ (2011) Metabolite-enabled eradication of bacterial persisters by aminoglycosides. Nature 473:216–220

20

Elen Louwagie et al.

131. Kwan BW, Chowdhury N, Wood TK (2015) Combatting bacterial infections by killing persister cells with mitomycin C. Environ Microbiol 17:4406–4414 132. Morones-Ramirez JR, Winkler JA, Spina CS, Collins JJ (2013) Silver enhances antibiotic activity against gram-negative bacteria. Sci Transl Med 5:190ra81 133. Starkey M, Lepine F, Maura D et al (2014) Identification of anti-virulence compounds that disrupt quorum-sensing regulated acute and persistent pathogenicity. PLoS Pathog 10:e1004321 134. Lebeaux D, Chauhan A, Le´toffe´ S et al (2014) pH-mediated potentiation of

aminoglycosides kills bacterial persisters and eradicates in vivo biofilms. J Infect Dis 210:1357–1366 135. Defraine V, Verstraete L, Van Bambeke F et al (2017) Antibacterial activity of 1-[(2,4-dichlorophenethyl)amino]-3-phenoxypropan-2-ol against antibiotic-resistant strains of diverse bacterial pathogens, biofilms and in pre-clinical infection models. Front Microbiol 8:2585 136. Stover CK, Warrener P, VanDevanter DR et al (2000) A small-molecule nitroimidazopyran drug candidate for the treatment of tuberculosis. Nature 405:962–966

Part II Quantification and Enrichment of Persister Cells

Chapter 2 Antibiotic Tolerance and Persistence Studied Throughout Bacterial Growth Phases Enea Maffei , Cinzia Fino , and Alexander Harms Abstract Antibiotic tolerance and persistence allow bacteria to survive lethal doses of antibiotic drugs in the absence of genetic resistance. Despite the urgent need to address these phenomena as a cause of clinical antibiotic treatment failure, studies on antibiotic tolerance and persistence are notorious for contradictory and inconsistent findings. Many of these problems are likely caused by differences in the methodology used to study antibiotic tolerance and persistence in the laboratory. Standardized experimental procedures would therefore greatly promote research in this field by facilitating the integrated analysis of results obtained by different research groups. Here, we present a robust and adaptable methodology to study antibiotic tolerance/persistence in broth cultures of Escherichia coli and Pseudomonas aeruginosa. The hallmark of this methodology is that the formation and disappearance of antibiotic-tolerant cells is recorded throughout all bacterial growth phases from lag after inoculation over exponential growth into early and then late stationary phase. In addition, all relevant experimental conditions are rigorously controlled to obtain highly reproducible results. We anticipate that this methodology will promote research on antibiotic tolerance and persistence by enabling a deeper view at the growth-dependent dynamics of this phenomenon and by contributing to the standardization or at least comparability of experimental procedures used in the field. Key words Antibiotic tolerance, Bacterial persistence, Bacterial resilience, Stationary phase, Growth media, Reproducibility, Escherichia coli, Pseudomonas aeruginosa

1

Introduction Since their initial observation in the 1940s by Bigger, bacterial antibiotic tolerance and persistence have drawn the attention of researchers and clinicians all over the world [1]. By definition, antibiotic tolerance is the ability of bacteria to survive nominally lethal concentrations of bactericidal antibiotics, while antibiotic persistence denotes the survival of a typically nongrowing, antibiotic-tolerant subpopulation among a heterogeneous population of otherwise sensitive, growing cells [2]. Though these phe-

Natalie Verstraeten and Jan Michiels (eds.), Bacterial Persistence: Methods and Protocols, Methods in Molecular Biology, vol. 2357, https://doi.org/10.1007/978-1-0716-1621-5_2, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2021

23

24

Enea Maffei et al.

nomena are thought to play major roles in antibiotic treatment failure and the frequent relapses of chronic infections [3, 4], their physiological basis and the underlying molecular mechanisms have remained largely unknown despite extensive research [5]. One major obstacle to progress in this field is that the published literature contains many contradictory or at least inconsistent findings [2, 5–8]. Dedicated research has suggested that these issues might be largely due to differences in the experimental methodologies used to assess antibiotic tolerance and persistence in the laboratory [2, 9–11]. In particular, the genetic background of the bacterial model organisms, their growth medium and culture conditions, and the inoculation method can have a large impact on the number of cells recovered after antibiotic treatment [9–13]. Recently, the field made an important effort to agree on common definitions of antibiotic tolerance and antibiotic persistence which included fundamental experimental guidelines on how to study these phenomena and how to distinguish them from antibiotic resistance [2]. However, no direct practical advice regarding experimental methodologies was included that could help researchers in the field understand their different results and guide newcomers around common pitfalls. The methodology presented in this chapter was developed for Escherichia coli and Pseudomonas aeruginosa, two major model organisms in the field, and largely relies on the rigorous control of all relevant experimental variables by, for example, using a fully defined growth medium to avoid comparing “apples, oranges and unknown fruit” [14]. Rather than artificially inducing antibiotic tolerance through some kind of bacteriostatic treatment [15–17], our methodology solely focuses on antibiotic-tolerant cells that form in response to cues inherent to the bacterial life cycle such as starvation or entry into stationary phase. In addition, our methodology provides a complete view on the dynamics of antibiotic tolerance by recording the formation and disappearance of antibiotic-tolerant cells throughout the bacterial growth phases from lag after inoculation over exponential growth into early and late stationary phase. This comprehensive view enables the comparison of mutant strains with different growth dynamics and can reveal phenotypes that affect tolerance or persistence only in certain growth stages [8, 18]. Conversely, the focus on single, defined growth time points for antibiotic tolerance assays is prone to comparing bacteria in different growth stages and blind to whether observed differences between strains are, for example, due to changes in the formation or in the disappearance of antibiotictolerant cells.

Antibiotic Tolerance Throughout Bacterial Growth Phases

2

25

Materials Prepare all solutions using autoclaved ultrapure water (Milli-Q grade) and store all reagents at room temperature in the dark unless indicated otherwise. All solutions, including the antibiotic stock solutions, should be sterilized by filtration through a 0.2 μm filter before use, unless indicated otherwise (see Notes 1 and 2). 1. 5 M9 salts solution: 33.9 g/L Na2HPO4, 15 g/L KH2PO4, 5 g/L NH4Cl, 2.5 g/L NaCl. Dissolve the powder in ca. 80% of the final volume of the solution, for example, in 800 mL of water if you are preparing 1 L of solution, in a large beaker. Mix well using a magnetic stirrer. Once all the powder is dissolved, fill up to the desired volume using a graduated cylinder. M9 salts can either be prepared from the individual components (see Note 3) or be bought as premixed powder (e.g., SigmaAldrich M6030). This solution can be stored for years in the dark. 2. 100 trace elements solution: 0.18 g/L ZnSO4·7H2O, 0.12 g/L CuCl2·2H2O, 0.12 g/L MnSO4·H2O, 0.18 g/L CoCl2·6H2O. Weigh the powders into a beaker and dissolve directly into the final volume of water. Store in the dark. 3. 100 mM FeCl3 solution: 16.22 g/L FeCl3. Dissolve the powder directly into the final volume of water and mix well. Store in the dark at 4  C. FeCl3 is corrosive, so special care should be taken and gloves as well as protective goggles should always be worn when manipulating it. 4. 1 M MgSO4 solution: 246.47 g/L MgSO4·7H2O. Dissolve the powder in the final volume of water and mix well. 5. 1 M CaCl2 solution: 147.01 g/L CaCl2·2H2O. Dissolve the powder in the final volume of water and mix well. 6. 2.2 M or 40% (w/v) D-glucose solution: 440 g/L D-glucose monohydrate (C6H12O6·H2O). Add the powder to ca. 70% of the final volume of water. Stir until everything is dissolved, subsequently fill up to the final volume using a measuring cylinder. Store at 4  C (see Note 4). 7. M9Glc or M9-based culture medium (50 mL): 38.75 mL of ultrapure water (sterilized), 10 mL of 5 M9 salts solution, 500 μL of 40% w/v D-glucose solution, 500 μL of 100 trace elements solution, 30 μL of 100 mM FeCl3 solution, 100 μL of 1 M MgSO4 solution, 5 μL of 1 M CaCl2 solution. Using sterile technique, mix all the substances in the order as listed to avoid precipitation. The final pH should be close to 7.2. Store at 4  C for up to 1 week. Final molar concentrations of all

26

Enea Maffei et al.

Table 1 Composition of M9Glc medium. This table summarizes the components needed to prepare M9Glc medium

Component

Stock solution

Volume (per 50 mL)

Final concentration

Ultrapure water

–

38.7 mL

–

5 M9 salts solution

5

10 mL

1

D-glucose solution

2.2 M (40% w/v D-glucose)

500 μL

2.2 mM (0.4% w/v D-glucose)

Trace elements solution

100

500 μL

1

FeCl3 solution

100 mM

30 μL

0.06 mM

MgSO4 solution

1M

100 μL

2 mM

CaCl2 solution

1M

5 μL

0.1 mM

components are listed in Table 1 (see also Note 5). This medium is referred to as “M9Glc medium” throughout the chapter. 8. PBS (phosphate-buffered saline) solution: 8 g/L NaCl, 0.2 g/ L KCl, 1.44 g/L Na2HPO4·2H2O, 0.24 g/L KH2PO4. Dissolve in 90% of the final volume of water and adjust the pH to 7.4 using 10 M NaOH. Fill up with water to the final volume and autoclave to sterilize. 9. LB agar plates: 10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl, 15 g/L agar, square petri dishes (12 cm  12 cm), round petri dishes (9.4 cm diameter). Mix all components and add water to the final volume. Sterilize by autoclaving. After sterilization, let the liquid LB agar cool down to ca. 60  C and subsequently pour the required number of plates by dispensing 45 mL of LB agar per square petri dish or 24 mL of LB agar per round petri dish. Leave the agar to solidify with the lid closed at room temperature. After 2 days, the plates are sufficiently dry for usage and can be stored in tightly closed plastic bags at 4  C for several weeks (see Notes 6 and 7). 10. Antibiotic stock solutions (50 mg/mL ampicillin; 1 mg/mL ciprofloxacin; 10 mg/mL tobramycin sulfate; 5 mg/mL meropenem trihydrate): Dissolve 53.14 mg ampicillin sodium salt; 1 mg ciprofloxacin; 10 mg tobramycin sulfate; or 5.7 mg meropenem trihydrate in 1 mL of ultrapure water each (see Note

Antibiotic Tolerance Throughout Bacterial Growth Phases

27

8). Ampicillin sodium salt, ciprofloxacin, and tobramycin sulfate stocks can be prepared in advance and stored at 20  C for several weeks, but multiple cycles of freezing and thawing should be avoided. Meropenem must be prepared freshly before each use due to low stability of the solution (see Notes 8–10). 11. Dilution plates: To prepare for serial dilutions, fill all wells in rows B-H of the required number of columns (one per sample) of a microtiter 96-well plate with 180 μL of PBS. To avoid evaporation, this should be done shortly before the serial dilutions are performed. Keep sterile. 12. Plastic test tubes: At least one sterile 1.5 mL test tube is needed per tested condition and time point. An excess of tubes should be sterilized by autoclaving before the experiment. 13. Flasks and glass tubes: One 50 mL Erlenmeyer flask per strain, one 100 mL Erlenmeyer flask per strain, and one cylindrical round-bottom glass culture tube with loose aluminium cap (15 mL nominal volume; 10  1.5 cm) per strain and condition are needed to determine levels of antibiotic-tolerant cells along the bacterial growth curve. One 50 mL Erlenmeyer flask per strain, one 500 mL Erlenmeyer flask per eight experimental conditions, and one cylindrical round-bottom glass culture tube with loose aluminium cap (15 mL nominal volume; 10  1.5 cm) per strain and condition are needed to perform the stationary-phase assay (see Note 11). 14. Graduated pipettes: To pipette and transfer larger volumes of media and culture, adequate sterile graduated pipettes should be used. 15. Bacterial strains: The protocols were developed for standard laboratory strains E. coli K-12 MG1655 F λ ilvG rfb-50 rph-1 (Coli Genetic Stock Center (CGSC) #6300) and P. aeruginosa PAO1 Δpel Δpsl [19] (see Note 12). 16. Cultivation and incubation: Bacterial liquid cultures should be agitated in a shaking incubator at 37  C at 170 rpm. Agar plates should be incubated without agitation at 37  C.

3

Methods In this part, the experimental procedures to determine antibiotic tolerance of E. coli K-12 MG1655 and P. aeruginosa PAO1 are described. In the first section, we describe the procedure to determine levels of antibiotic-tolerant cells of a bacterial culture followed from lag to early stationary phase. In the second section, we describe the procedure to determine levels of antibiotic-tolerant cells in deep stationary phase.

28

Enea Maffei et al.

Table 2 Materials per strain tested. The table summarises the materials and media needed per strain for the experiment “Determination of antibiotic tolerance along the growth curve” (Subheading 3.1) M9Glc medium

ca. 100 mL

Phosphate-buffered saline (PBS) solution

ca. 70 mL

Glass culture tube

8

LB-agar square petri dish

2

100 mL Erlenmeyer flask

8

1.5 mL plastic test tube

16

96-Well microtiter plate

2

5 mL graduated pipette

16

3.1 Determination of Antibiotic Tolerance Along the Growth Curve 3.1.1 Day 0: Collection of Materials and Inoculation of the Overnight Culture 3.1.2 Day 1: Subculturing

1. Prepare or collect all the materials needed for the assay (step 1 in Subheading 3.1.3 and following) (Table 2). 2. Pick an isolated bacterial colony and use it to inoculate 5 mL of M9Glc medium in a 50 mL Erlenmeyer flask. Incubate the preculture for 24 h at 37  C shaking (170 rpm) (see Notes 11–13).

1. Dispense 10 mL of M9Glc medium into one 100 mL flask and let the medium reach room temperature. 2. Dilute the preculture 1:100 into the flask prepared in the previous step. Incubate this secondary preculture at 37  C shaking (170 rpm) for 24 h (see Notes 11 and 12).

3.1.3 Day 2: Antibiotic Tolerance Assay

1. Switch on the shaking incubator (37  C, 170 rpm) at least 30 min before the start of the experiment to let it heat up. In the meantime, take the M9Glc medium out of the fridge, aliquot it into the glass flasks using sterile techniques, and let the medium reach room temperature. While waiting for the medium to warm up, the tubes needed for the assay can be labeled and other preparatory work can be done. 2. Dilute the preculture 1:100 into 10 mL of fresh M9Glc medium in a 100 mL Erlenmeyer flask and culture both the subculture and the preculture at 37  C, 170 rpm. Repeat this subculturing step after 2 h and then six more times in a way that, eventually, eight cultures inoculated 8 h, 6 h, 5 h, 4 h, 3 h, 2 h, and 1 h before the last one are available (see Notes 11, 12 and 14).

Antibiotic Tolerance Throughout Bacterial Growth Phases

29

Fig. 1 Serial dilutions and spotting. This figure is a schematic representation of the serial dilutions performed in a 96-well microtiter plate (left) and the transfer of all dilutions onto a dry LB agar plate in a square petri dish for colony outgrowth (right). Briefly, 100 μL of each sample are transferred into the wells of row A after the wells in the other rows (B–H) have been filled with each 180 μL of sterile PBS. Serial dilutions (1:10) are performed with a multichannel pipet by transferring 20 μL from one row to the next (as indicated by arrows; left), mixing, discarding the tips, and then repeating this process until row H (dilution 10 8) is reached. Subsequently, 10 μL spots of all dilutions are transferred onto the LB agar plate from the highest dilutions to the undiluted samples using a single set of tips (rows H–A, arrows 1–8)

3. Determine the number of colony forming units (CFU/mL) in the eight cultures first directly to assess the total number of bacteria (Subheading 3.1.3 steps 4–6) and, subsequently, after antibiotic treatment (Subheading 3.1.3 steps 7–10) to assess the number of survivors (Subheading 3.1.4 and following). 4. Transfer 100 μL of each culture into the wells of row A of a 96-well microtiter plate previously filled with 180 μL PBS. Perform 1:10 serial dilutions with the multichannel pipette by transferring 20 μL of the cultures from one row to the next containing PBS. Mix well and change tips after every dilution (Fig. 1, left). 5. Use a multichannel pipet to transfer 10 μL spots of all dilutions from the 96-well plate onto a dry LB agar square petri dish (see Note 6; Fig. 1, right). 6. Wait for the spots to dry with open lid (e.g., next to a Bunsen burner or under a flow hood). Once the spots have dried, incubate the plate at 37  C for 16–24 h before counting. 7. For the antibiotic treatment, transfer 3 mL of each subculture from the flask into a glass tube and challenge with antibiotics at the desired final concentration. We typically treat with 100 μg/ mL ampicillin (E. coli) or 12 μg/mL meropenem (P. aeruginosa), 10 μg/mL ciprofloxacin, and 40 μg/mL tobramycin sulfate. Place the eight tubes in the incubator (37  C, 170 rpm) for 5 h (see Notes 11, 15 and 16).

30

Enea Maffei et al.

Fig. 2 Colony counts after CFU regrowth. This figure is a schematic representation of plates from a typical experiment of treatment of bacteria along the growth curve after colonies have grown for 24–48 h. Regions containing between 10 and 100 colonies are highlighted by a violet circle and should be used for CFU/mL calculation

8. After 5 h of antibiotic challenge, transfer 1.5 mL of each culture to a sterile 1.5 mL plastic tube and pellet the cells by centrifugation for 2 min at 18,000  g. Remove the supernatant carefully and wash each pellet with 1 mL of sterile PBS. Start by washing the samples with the lowest incubation times and pay attention to the small, barely visible pellets that result from low-density cultures (see Notes 17 and 18). Repeat the centrifugation, remove the supernatant, and finally resuspend the pellets in 100 μL of sterile PBS. 9. Dilute and spot these samples obtained from the antibiotictreated samples as described in Subheading 3.1.3 steps 4 and 5 (Fig. 1). 10. Let the spots dry and incubate the plate at 37  C for at least 16 h but monitor colony formation for 48 h (see Fig. 2 and Notes 19 and 20). 3.1.4 Day 3: CFU Counting

1. After 16–24 h of incubation, select the spots that contain between 10 and 100 bacterial colonies. Count the CFU per spot (CFU/spot). Subsequently, put the plates back for an additional incubation of ca. 24 h at 37  C (see Fig. 2 and Note 19).

3.1.5 Day 4: CFU Counting and Plotting

1. Check the plates from day 3 and, if necessary, update the CFU counts. Subsequently, calculate the CFU/mL values for all samples before and after antibiotic treatment.

Antibiotic Tolerance Throughout Bacterial Growth Phases

31

Fig. 3 Antibiotic tolerance of E. coli and P. aeruginosa along the growth phases. (a) Levels of E. coli K-12 MG1655 cells tolerant to 100 μg/mL ampicillin (light blue), 10 μg/mL ciprofloxacin (green), or 40 μg/mL tobramycin sulfate (red) were recorded throughout the bacterial growth phases from inoculation to early stationary phase. (b) Levels of P. aeruginosa PAO1 Δpel Δpsl cells tolerant to 12 μg/mL meropenem (dark blue), 10 μg/mL ciprofloxacin (green), or 40 μg/mL tobramycin sulfate (red) were recorded throughout the bacterial growth phases from inoculation to early stationary phase. Results of one representative experiment are shown. The limit of detection (100 CFU/mL) is indicated by the upper edge of the gray bar

2. Optional: Calculate the fraction of antibiotic-tolerant cells at each time point as the ratio of CFU/mL after and before antibiotic treatment. 3. To obtain independent biological replicates, the experiment should be performed at least three times on different days. When sufficient biological replicates have been performed, the data can be averaged and the standard error of the mean can be calculated. To calculate the average, transform all the CFU/mL values into their base 10 logarithm. Consequently, calculate the standard error of the mean on the individual log-transformed values. 4. Plot the recovered CFU/mL for each time point before and after antibiotic treatment on a base 10 logarithmic scale against time of growth before treatment (Fig. 3). If enough replicates have been performed, plot the mean values instead and include the respective error values as well. Note that these curves per se only report on antibiotic tolerance and that the inference of

32

Enea Maffei et al.

Table 3 Materials per strain tested, The table summarises the materials and media needed per strain for the experiment “Determination of antibiotic tolerance in stationary phase” (Subheading 3.2) M9Glc medium

ca. 60 mL

Phosphate-buffered saline (PBS) solution

ca. 50 mL

Glass culture tube

8

LB-agar square petri dish

6

100 mL Erlenmeyer flask

2

500 mL Erlenmeyer flask

2

1.5 mL plastic test tube

40

96-Well microtiter plate

6

5 mL graduated pipette

2

antibiotic persistence requires additional time kill curves (see Note 20). The results and their interpretation are briefly discussed in Note 21. 3.2 Determination of Antibiotic Tolerance in Stationary Phase 3.2.1 Day 0: Preparation of Media and Cultures

3.2.2 Day 1: Subculturing

1. Prepare or collect all the materials needed for the assay (Subheading 3.2.3 step 2 and following) (Table 3). 2. Pick an isolated bacterial colony and use it to inoculate 5 mL of M9Glc medium in a 50 mL Erlenmeyer flask. Incubate the preculture for 24 h at 37  C shaking (170 rpm) (see Notes 11–13). 1. Dispense 50 mL of M9Glc medium into one 500 mL Erlenmeyer flask and let the medium warm up to room temperature. 2. Dilute the preculture 1:100 into the flask prepared in the previous step. Incubate the subculture at 37  C shaking (170 rpm) for 48 h (see Notes 11 and 12).

3.2.3 Day 3: Stationary-Phase Antibiotic Tolerance Assay

1. For the antibiotic treatment, transfer 5 mL of each culture from the Erlenmeyer flask into a glass tube and challenge with antibiotics at the desired final concentration. We typically treat with 100 μg/mL ampicillin (E. coli) or 12 μg/mL meropenem (P. aeruginosa), 10 μg/mL ciprofloxacin, and 40 μg/mL tobramycin sulfate. Place the tubes in a shaking incubator (37  C/170 rpm) (see Notes 11, 15 and 16). Always include one untreated sample as a control for bacterial stationary-phase viability. 2. To determine viable cell counts, withdraw 200 μL aliquots from each glass tube at different time points, typically after 0 h (i.e., before addition of the antibiotics), 1 h, 3 h, 7 h,

Antibiotic Tolerance Throughout Bacterial Growth Phases

33

24 h, as well as 48 h, and transfer into a 1.5 mL tube. Centrifuge each tube for 2 min at 18,000  g and wash once with 200 μL of sterile PBS to remove excess antibiotic. 3. Transfer 100 μL of each culture into the wells of row A of a 96-well microtiter plate previously filled with 180 μL PBS. Perform 1:10 serial dilutions with the multichannel pipette by transferring 20 μL of the cultures from one row to the next containing sterile PBS. Mix well and change tips after every dilution (similar as shown in Fig. 1, left). 4. Use a multichannel pipet to transfer 10 μL spots of all dilutions from the 96-well plate onto a dry LB agar square petri dish (see Note 6; Fig. 1, right). 5. Wait for the spots to dry with open lid (e.g., next to a Bunsen burner or under a flow hood). Once the spots have dried, incubate the plate at 37  C for 16–24 h. 3.2.4 Day 4: CFU Counting

1. After 16–24 h of incubation, select the spots that contain between 10 and 100 bacterial colonies. Count the CFU per spot (CFU/spot). Subsequently put the plates back for an additional incubation of ca. 24 h at 37  C (see Fig. 2 and Note 19).

3.2.5 Day 5: CFU Counting and Plotting

1. Check the plates from day 4 and, if necessary, update the CFU counts. Subsequently, calculate the CFU/mL values for all samples before and after antibiotic treatment. 2. To obtain independent biological replicates, the experiment should be performed at least three times on different days. Plot the recovered CFU/mL for each time point on a base 10 logarithmic scale against time of treatment (Fig. 4). If enough replicates have been performed, plot the mean values instead and include the respective error values as well. The results and their interpretation are briefly discussed in Note 22.

4

Notes 1. Sterilization by filtration should always be preferred since autoclaving can alter the chemical composition of solutions. This is particularly important for solutions containing complex molecules such as glucose which can degrade with heat [20]. 2. All experimental steps, from the preparation of the media to the antibiotic killing assays, should be performed according to the rules of good laboratory practice [21]. 3. If preparing the M9 salts solution from the individual components, be aware of water of hydration inside the salt crystals (e.g., disodium hydrogen phosphate is usually available as

34

Enea Maffei et al.

Fig. 4 Antibiotic persistence of E. coli and P. aeruginosa in stationary phase. (a) A stationary-phase culture of E. coli K-12 MG1655 was treated with 100 μg/mL ampicillin (light blue), 10 μg/mL ciprofloxacin (green), or 40 μg/mL tobramycin sulfate (red) and bacterial survival was recorded over time. (b) A stationary-phase culture of P. aeruginosa Δpel Δpsl was treated with 12 μg/mL meropenem (dark blue), 10 μg/mL ciprofloxacin (green), or 40 μg/mL tobramycin sulfate (red) and bacterial survival was recorded over time. Data points represent average values of three biological replicates and error bars show the standard error of the mean. The limit of detection is 100 CFU/mL (not shown)

Na2HPO4·7H2O) and adapt the recipe accordingly to preserve molarities. The recipe described in this chapter corresponds to the quantities used by Sigma-Aldrich to prepare their variant of M9 salts powder (Sigma-Aldrich M6030). Note that there are some minor differences between M9 salts recipes that are circulating in the community, for example, between the recipe of Sigma-Aldrich and the one formulated by Cold Spring Harbor Protocols [22]. 4. Dissolving the glucose may take up to 1 h. It is not recommendable to add water to dissolve the glucose powder more quickly, as the volume of the solution increases considerably when the D-glucose is dissolving. Instead, heating the solution gradually up to 40  C can help dissolving the powder more quickly. 5. The main benefit of using a chemically defined medium prepared with ultrapure water is that the same cellular physiology can be obtained in every experiment and among different

Antibiotic Tolerance Throughout Bacterial Growth Phases

35

laboratories. This is often not the case for complex media that may suffer from batch-to-batch variation due to poorly defined ingredients (e.g., tryptone and yeast extract) or due to degradation of some components if stored or prepared improperly [8, 12]. The trace element supplement is adapted from previous work by Gerosa et al [23]. Unless strictly necessary, we do not suggest to supplement the medium with thiamine (as it is done in many M9 medium recipes including the one of Gerosa et al [23]) in order to avoid that this compound might be used as an alternative carbon or nitrogen source. To assess antibiotic tolerance of thiamine auxotrophs with the present protocol, it is enough to supplement the medium with thiamine to a final concentration of 2.8 μM (as described in reference (23)). Any other auxotrophies can be complemented analogously. 6. Drying the plates at room temperature for up to 48 h (not wrapped in a plastic bag) generates a more hygroscopic surface which avoids that drops of bacterial samples spotted closely flow into each other and merge. Furthermore, a dry agar plate will cause bacterial samples to soak into the agar more quickly. 7. Previous experience of the authors showed that P. aeruginosa is very sensitive to the brand of agar used to prepare LB agar plates. Depending on the brand and the overall concentration of agar used, the colony morphology of P. aeruginosa and surface-dependent behaviors like swarming motility can be very different. In this work, AppliChem A0949 agar was used. If it is not possible to use this brand, it might be worth to test a few different brands or concentrations (13–17 g/L) to see which one gives an optimal colony size after 24 h of incubation at 37  C (3–5 mm). Notably, E. coli seems to be not visibly affected by agar brand and concentration. 8. Antibiotic powders are often provided in the form of a salt (e.g., ampicillin sodium salt). In these cases, antibiotic stocks should be prepared so that the final concentration of the stock corresponds to the actual antibiotic concentration and not to the concentration of its salt. When preparing ciprofloxacin antibiotic stock from pure ciprofloxacin powder, a few drops of 1 M HCl must be added to help complete dissolution of the powder in ultrapure water. Alternatively, it is also possible to purchase ciprofloxacin hydrochloride monohydrate which is readily soluble in water. 9. P. aeruginosa is intrinsically resistant to ampicillin due to an inducible AmpC β-lactamase [24]. Therefore, a different β-lactam antibiotic must be used to assess the tolerance of P. aeruginosa to β-lactams. We propose to use meropenem,

36

Enea Maffei et al.

but also other β-lactams can be used such as imipenem, piperacillin, or cefepime. 10. Antibiotic tolerance and persistence can either be confined to single drugs or affect multiple antibiotics, and the levels of antibiotic-tolerant cells can vary widely among different strains of the same species [25–27]. In order to obtain a complete picture, we therefore recommend to always separately assess tolerance to antibiotics with different modes of action such as β-lactams, fluoroquinolones, and aminoglycosides that are commonly used in the field (see also Figs. 3 and 4). 11. The proper aeration of a culture during long incubation times is essential to avoid uncontrolled changes of bacterial physiology that could affect outcome of the antibiotic tolerance assay. Therefore, make sure to cultivate the bacteria in a volume of medium that is around ten times less than the nominal fill volume of the Erlenmeyer flasks. Shaking tubes and flasks at 170 rpm is suitable to achieve both proper aeration and to prevent sedimentation of the bacteria. Whenever having a culture in a non–self-sustaining container (e.g., a glass culture tube), this should be inclined at ca. 45 to ensure optimal aeration. The inclination should be kept constant for the whole duration of the assay. 12. P. aeruginosa can form small aggregates in liquid batch culture that exhibit biofilm-like characteristics including, for example, increased antibiotic tolerance and therefore distort the results of antibiotic treatment assays. However, the formation of these aggregates can be largely avoided by using a mutant deficient in the production of Pel and Psl exopolysaccharides as well as by using precultures rather than direct inoculation from a single colony or cryostocks [9]. Nevertheless, visible aggregates of P. aeruginosa can transiently form during the precultures and subcultures but disperse again upon starvation [28]. Therefore, take care while pipetting to not include visible bacterial aggregates in the inocula for subcultures. 13. The colonies should be inoculated from a freshly streaked plate that is not older than 1 week. Alternatively, the preculture can also be inoculated directly from a cryostock. It is advisable to keep the inoculation method consistent among experiments, as it may have an influence on the results [9, 11]. 14. It is recommended to prepare, collect, and label wherever necessary all the materials that are required throughout the steps prior to experimentation. 15. Measure the antibiotic minimum inhibitory concentration (MIC) by broth dilution assays in 96-well microtiter plates to make sure that the concentration used for the antibiotic

Antibiotic Tolerance Throughout Bacterial Growth Phases

37

Table 4 MIC values of E. coli K-12 MG1655 and P. aeruginosa PAO1 Δpel Δpsl in M9Glc medium. This table presents the minimum inhibitory concentration (MIC) values of E. coli K-12 MG1655 and P. aeruginosa PAO1 Δpel Δpsl in M9Glc as determined by the authors for the antibiotics used in this study. The use of different β-lactam antibiotics for E. coli and P. aeruginosa is addressed in Note 9. The values reported are the result of three independent biological replicates. N.D. not determined Tobramycin, μg/mL

Ciprofloxacin, μg/mL

Ampicillin, μg/mL

Meropenem, μg/mL

Escherichia coli K-12 MG1655

0.25

0.015

3

N.D.

Pseudomonas aeruginosa PAO1 Δpel Δpsl

1

0.06

N.D.

0.5

tolerance assay is considerably higher than this value [29]. If in doubt about the optimal antibiotic concentration for tolerance and persistence assays, test a range of concentrations and choose one that is so high that the killing rate does not increase anymore with increasing drug concentration. When comparing the antibiotic killing dynamics of different bacterial mutants or strains, measure the MIC of each of them. If they are very different, consider for your experiments that not only differences in antibiotic tolerance but also different intrinsic drug resistance (as expressed by MIC) can change the dynamics of bacterial killing in antibiotic treatment assays. See Table 4 for reference MIC values of E. coli K-12 MG1655 and P. aeruginosa PAO1. 16. Treating with 10 μg/mL ciprofloxacin, a very high concentration (ca. 1000 MIC), is necessary to avoid artefacts caused by secondary killing of antibiotic-tolerant bacteria due to the induction of prophages [8]. 17. To increase accuracy when pipetting small pellets and to reduce the risk of disturbing the pellet, one can fit a regular 200 μL pipette tip on top of a 1000 μL tip. Like this, the large volume pipetted during washing can still be removed efficiently but the risk to wash out the small and sometimes invisible pellets at the bottom of the tube is reduced. 18. The pellets from samples of late-exponential-phase cultures treated with β-lactams are often difficult to homogeneously resuspend due to fulminant bacterial lysis and aggregation of cell debris. Vortexing the tubes for 15–30 s and pipetting the pellets up and down a few times are sufficient to reproducibly release the surviving bacteria into the PBS, because most of the hard pellets itself is merely composed of dead cell debris.

38

Enea Maffei et al.

19. Bacteria treated with ciprofloxacin or tobramycin can take more than 36 h to form visible colonies on LB agar plates, in particular in case of P. aeruginosa. In order to include all bacterial survivors, it is therefore important to incubate the LB agar plates for 40–48 h in total at 37  C before the final CFU counts are recorded. 20. The hallmark of antibiotic persistence is a time kill curve with a biphasic trajectory [2]. In cases where it is necessary to distinguish between antibiotic tolerance and persistence, it is therefore important to verify biphasic killing by determining viable CFU/mL not only after 5 h of drug treatment but also after, for example, 1 h, 3 h, and 7 h in order to record the dynamics of bacterial killing (like in a regular time kill curve). Previous work showed that bacterial killing under conditions similar to those presented in this protocol is largely biphasic [8, 18]. 21. The results presented in Fig. 3 highlight a few important aspects. First, the overall shape of the tolerance curves along the growth phases are very similar for both E. coli K-12 MG1655 and P. aeruginosa PAO1 Δpel Δpsl, suggesting that the underlying phenomena causing antibiotic tolerance are similar irrespective of the model organism investigated (similar to previous results [30]). Second, it is well visible how the initially higher tolerance levels (up to 2–3 h of cultivation) are a consequence of stationary phase carryover and drop as more bacteria exit lag phase [8]. Third, once bacteria start to successively enter stationary phase, the levels of tolerant cells increase again. For both organisms, tobramycin tolerance is only observed in stationary phase. The dynamics of ampicillin and ciprofloxacin tolerance are largely parallel except for the exponential phase of E. coli (ca. 3–6 h after subculturing) where the stark decrease in ampicillin-tolerant cells is not mirrored by the levels of ciprofloxacin-tolerant cells. 22. The results presented in Fig. 4 reveal a massive presence of antibiotic-tolerant cells for both E. coli K-12 MG1655 and P. aeruginosa PAO1 Δpel Δpsl in the stationary phase. The biphasic kill curve resulting from ciprofloxacin treatment reveals a large fraction of persisters and at the same time indicates that most bacteria in culture are still active in DNA processing (as a prerequisite for gyrase poisoning by fluoroquinolones). Whereas the complete tobramycin tolerance of both organisms can be explained by reduced drug uptake due to lowered membrane potential, the total tolerance observed against β-lactams suggests absence of actively dividing cells in the later stages of stationary phase under these conditions [11, 31, 32].

Antibiotic Tolerance Throughout Bacterial Growth Phases

39

Acknowledgments The authors are grateful to Prof. Urs Jenal, Dr. Alexander Klotz, Dr. Pablo Manfredi, and Margo van Berkum for useful discussions about optimal medium composition and growth conditions for antibiotic treatment experiments with E. coli and P. aeruginosa. Furthermore, the authors thank Prof. Urs Jenal for providing strain P. aeruginosa PAO1 Δpel Δpsl. This work was supported by the Swiss National Science Foundation (SNSF) Ambizione Fellowship PZ00P3_180085 and SNSF National Centre of Competence in Research (NCCR) AntiResist. References 1. Bigger JW (1944) Treatment of staphylococcal infecctions with penicillin by intermittent sterilisation. Lancet 244:497–500 2. Balaban NQ, Helaine S, Lewis K et al (2019) Definitions and guidelines for research on antibiotic persistence. Nat Rev Microbiol 17:441–448 3. Mulcahy LR, Burns JL, Lory S, Lewis K (2010) Emergence of Pseudomonas aeruginosa strains producing high levels of persister cells in patients with cystic fibrosis. J Bacteriol 192:6191–6199 4. Lewis K, Manuse S (2019) Persister formation and antibiotic tolerance of chronic infections. In: Lewis K (ed) Persister cells and infectious disease. Springer, New York 5. Kaldalu N, Hauryliuk V, Tenson T (2016) Persisters-as elusive as ever. Appl Microbiol Biotechnol 100:6545–6553 6. Pontes MH, Groisman EA (2019) Slow growth determines nonheritable antibiotic resistance in Salmonella enterica. Sci Signal 12:eaax3938 7. Goormaghtigh F, Fraikin N, Putrinsˇ M et al (2018) Reassessing the role of type II toxinantitoxin systems in formation of Escherichia coli type II persister cells. mBio 9: e00640–e00618 8. Harms A, Fino C, Sørensen MA et al (2017) Prophages and growth dynamics confound experimental results with antibiotic-tolerant persister cells. mBio 8:1–18 9. Kragh KN, Alhede M, Rybtke M et al (2017) The inoculation method could impact the outcome of microbiological experiments. Appl Environ Microbiol 84:e02264–e02217 10. Heinemann M, Basan M, Sauer U (2020) Implications of initial physiological conditions for bacterial adaptation to changing environments. Mol Syst Biol 16:e9965

˜ ers A, Kaldalu N, Tenson T 11. Luidalepp H, Jo (2011) Age of inoculum strongly influences persister frequency and can mask effects of mutations implicated in altered persistence. J Bacteriol 193:3598–3605 12. Sezonov G, Joseleau-Petit D, D’Ari R (2007) Escherichia coli physiology in Luria-Bertani broth. J Bacteriol 189:8746–8749 13. Goormaghtigh F, Van Melderen L (2016) Optimized method for measuring persistence in Escherichia coli with improved reproducibility. In: Michiels J, Fauvart M (eds) Bacterial persistence: methods and protocols. Springer, New York, NY 14. Neidhardt FC (2006) Apples, oranges and unknown fruit. Nat Rev Microbiol 4:876–876 15. Kwan BW, Valenta JA, Benedik MJ, Wood TK (2013) Arrested protein synthesis increases persister-like cell formation. Antimicrob Agents Chemother 57:1468–1473 16. Hong SH, Wang X, O’Connor HF et al (2012) Bacterial persistence increases as environmental fitness decreases. Microb Biotechnol 5:509–522 17. Va´zquez-Laslop N, Lee H, Neyfakh AA (2006) Increased persistence in Escherichia coli caused by controlled expression of toxins or other unrelated proteins. J Bacteriol 188:3494–3497 18. Fino C, Vestergaard M, Ingmer H et al (2020) PasT of Escherichia coli sustains antibiotic tolerance and aerobic respiration as a bacterial homolog of mitochondrial Coq10. Microbiology 9:e1064 19. Broder UN, Jaeger T, Jenal U (2016) LadS is a calcium-responsive kinase that induces acuteto-chronic virulence switch in Pseudomonas aeruginosa. Nat Microbiol 2:1–11 20. Wang X-J, Hsiao K-C (1995) Sugar degradation during autoclaving: effects of duration and

40

Enea Maffei et al.

solution volume on breakdown of glucose. Physiol Plant 94:415–418 21. Ridley R (2009) Handbook good laboratory practice (GLP): quality practices for regulated non-clinical research and development, special programme for research & training in tropical diseases (TDR), 2nd edn. WHO, Geneva 22. Cold Spring Harbor Protocols (2009) M9 salt solution. https://doi.org/10.1101/pdb. rec11973. Accessed 19 Oct 2020 23. Gerosa L, Kochanowski K, Heinemann M, Sauer U (2013) Dissecting specific and global transcriptional regulation of bacterial gene expression. Mol Syst Biol 9:658 24. Livermore DM (1995) Lactamases in laboratory and clinical resistance. Clin Microbiol Rev 8:557–584 25. Hofsteenge N, Van Nimwegen E, Silander OK (2013) Quantitative analysis of persister fractions suggests different mechanisms of formation among environmental isolates of E. coli. BMC Microbiol 13:25 26. Harms A, Maisonneuve E, Gerdes K (2016) Mechanisms of bacterial persistence during stress and antibiotic exposure. Science 354: aaf4268

27. Lewis K (2019) Persister cells and infectious disease. Springer, New York, NY 28. Schleheck D, Barraud N, Klebensberger J et al (2009) Pseudomonas aeruginosa PAO1 preferentially grows as aggregates in liquid batch cultures and disperses upon starvation. PLoS One 4:e5513 29. Wiegand I, Hilpert K, Hancock REW (2008) Agar and broth dilution methods to determine the minimal inhibitory concentration (MIC) of antimicrobial substances. Nat Protoc 3:163–175 30. Keren I, Kaldalu N, Spoering A et al (2004) Persister cells and tolerance to antimicrobials. FEMS Microbiol Lett 230:13–18 31. Damper PD, Epstein W (1981) Role of the membrane potential in bacterial resistance to aminoglycoside antibiotics. Antimicrob Agents Chemother 20:803–808 32. Tuomanen E, Cozens R, Tosch W et al (1986) The rate of killing of Escherichia coli by β-lactam antibiotics is strictly proportional to the rate of bacterial growth. J Gen Microbiol 132:1297–1304

Chapter 3 A Robust Method for Generating, Quantifying, and Testing Large Numbers of Escherichia coli Persisters Silke R. Vedelaar, Jakub L. Radzikowski, and Matthias Heinemann Abstract Bacteria can exhibit phenotypes that render them tolerant against antibiotics. However, often only a few cells of a bacterial population show the so-called persister phenotype, which makes it difficult to study this health-threatening phenotype. We recently found that certain abrupt nutrient shifts generate Escherichia coli populations that consist almost entirely of antibiotic-tolerant cells. These nearly homogeneous persister cell populations enable assessment with population-averaging experimental methods, such as highthroughput methods. In this chapter, we provide a detailed protocol for generating a large fraction of tolerant cells using the nutrient-switch approach. Furthermore, we describe how to determine the fraction of cells that enter the tolerant state upon a sudden nutrient shift and we provide a new way to assess antibiotic tolerance using flow cytometry. We envision that these methods will facilitate research into the important and exciting phenotype of bacterial persister cells. Key words Antibiotic tolerance, Flow cytometry, Nutrient shift, Tolerant cells, Persister cells, Escherichia coli

1

Introduction Bacterial persistence is defined as the occurrence of cells within a population that are transiently tolerant against antibiotics without carrying a mutation conferring genetic resistance [1]. These antibiotic-tolerant cells have been suggested to be responsible for recurrent infections [2]. Tolerant cells can be formed stochastically in exponentially growing cultures [3]. Activation of toxin–antitoxin (TA) modules [2] and deletion of certain metabolic genes [4] has also been shown to increase the number of cells entering this

Electronic Supplementary Material: The online version of this chapter (https://doi.org/10.1007/978-10716-1621-5_3) contains supplementary material, which is available to authorized users. Natalie Verstraeten and Jan Michiels (eds.), Bacterial Persistence: Methods and Protocols, Methods in Molecular Biology, vol. 2357, https://doi.org/10.1007/978-1-0716-1621-5_3, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2021

41

42

Silke R. Vedelaar et al.

phenotype. Certain environmental perturbations can also induce the fraction of tolerant cells in a population, such as the entry into stationary phase [5] or certain sudden nutrient shifts, in which case almost all cells in a population enter the tolerant state [6, 7]. The fact that certain nutrient shifts, for example, the one from glucose to fumarate in E. coli [8], can force almost all the cells in a population into the tolerant state offers new research opportunities to investigate the molecular basis of tolerance and persistence. While the investigation of the few stochastically occurring persisters in exponentially growing cultures and the very heterogeneous stationary-phase cultures require single-cell analyses or cell-sorting approaches, a population that consists almost entirely of tolerant cells allows the use of population-level high-throughput analyses. For example, by exploiting sudden nutrient shifts to generate almost homogeneous populations of tolerant cells, it was found that tolerant cells have increased ppGpp levels, are metabolically active with a metabolism geared toward energy generation and catabolism, and exhibit a proteome characterized by the σSmediated stress response [8]. Despite the now enabled use of omics techniques for the study of bacterial persistence, experiments to assess how many cells are tolerant at which level of antibiotic exposure are still necessary. Here, in most cases, classical plating assays using different dilutions are performed to determine the colony-forming units (e.g., [5, 6]). However, this technique is very laborious and it has a high experiment-to-experiment variability requiring many plates to obtain statistically sound results [9]. Moreover, with plating fewer cells are recovered from their dormancy compared to recovery in liquid medium [10]. To this end, we recently introduced an alternative method to assess antibiotic tolerance with flow cytometry [8]. This method can be applied to populations solely consisting of dormant cells or to heterogeneous populations with dormant and growing cells. In this method, single cells’ regrowth is assessed via temporal dilution of a fluorescent signal, which can originate from a stained membrane [11] or GFP expression [12]. In the current chapter we provide a detailed nutrient-shiftbased protocol to generate large fractions of tolerant cells, which enables the study of tolerant cells on the population level. Furthermore, we illustrate a method to fluorescently stain the membrane of cells, which, together with flow cytometry and a Matlab script, allows to determine the fraction of cells that enter the tolerant state upon a sudden nutrient shift. Finally, we describe how the membrane-staining procedure together with flow cytometry can be used to assess tolerance.

Nutrient Shift to Generate Antibiotic-Tolerant Cells

2

43

Materials Prepare all media using demi water.

2.1 Generation of Large Fractions of Tolerant Cells

1. LB agar plates: Prepare LB medium by adding tryptone (to a final concentration of 1% w/v), yeast extract (0.5% w/v) and sodium chloride (1% w/v) to a bottle with demi water. Add agar to a final concentration of 1.5% (w/v) and mix well. Autoclave for 15 min at 121
C and let cool down to ~55
C. Add antibiotics, if needed, and pour ~20 mL LB-agar per 10 cm petri dish (see Note 1). Store plates sealed with Parafilm at 4
C with the agar side up. LB-plates containing antibiotics can be stored for up to 1 month. 2. M9 minimal medium: Prepare a base salt solution (211 mM Na2HPO4, 110 mM KH2PO4, 42.8 mM NaCl, 56.7 mM (NH4)2SO4), a solution with trace elements (0.63 mM ZnSO4, 0.7 mM CuCl2, 0.71 mM MnSO4, 0.76 mM CoCl2), a solution with 0.1 M CaCl2, one with 1 M MgSO4, one with 1.4 mM thiamine-HCl, and one with 0.1 M FeCl3. Autoclave the solutions with base salts, CaCl2, MgSO4, and trace elements, and store them at room temperature. Filter sterilize the thiamine and FeCl3 solutions using a 0.22 μm PES filter (see Note 2) and store them at 4
C. Stock solutions are sterilized so they can be stored for up to half a year. For 1 L M9 minimal medium, add 200 mL base salts, 700 mL water, 10 mL trace elements, 1 mL CaCl2 solution, 1 mL MgSO4 solution, 0.6 mL FeCl3 solution and 2 mL thiamine solution. Fill up to 1 L with water and filter sterilize the resulting solution using a PES bottle top filter into an autoclaved 1 L bottle. Store at 4
C for up to 1 month. The required carbon source is added just before the medium is used for cultivations. 3. 250 g/L glucose stock solution: For 100 mL, weigh 27.5 g D-glucose monohydrate and dissolve in 90 mL water. Adjust the pH to 7 using NaOH and fill up to 100 mL with water. Filter-sterilize the resulting solution using a PES bottle top filter into an autoclaved 100 mL bottle. Store at room temperature for up to 1 month. 4. 100 g/L fumarate stock solution: For 100 mL, weigh 14 g sodium fumarate and dissolve in 90 mL water. Adjust the pH to 7 using HCl and fill up to 100 mL with water. Filter-sterilize the resulting solution using a PES bottle top filter into an autoclaved 100 mL bottle. Store at room temperature for up to 1 month. 5. M9 minimal medium supplemented with glucose to a final concentration of 5 g/L for immediate use for cultivation:

44

Silke R. Vedelaar et al.

Add 1 mL glucose stock solution (250 g/L) to 49 mL minimal medium in a 500 mL Erlenmeyer flask. Preheat and pre-aerate at 37
C with shaking at 300 rpm before inoculation. 6. M9 minimal medium supplemented with fumarate to a final concentration of 2 g/L for immediate use for cultivation: Add 1 mL fumarate stock solution (100 g/L) to 49 mL minimal medium in a 500 mL Erlenmeyer flask. Preheat and pre-aerate at 37
C with shaking at 300 rpm before inoculation. 7. Flow cytometer: Accuri C6, BD Biosciences or equivalent. 2.2 Identification and Quantification of Tolerant Cells Using Membrane Staining and Flow Cytometry

1. M9 minimal medium without carbon source, ice-cold. 2. M9 minimal medium with 1% (w/v) bovine serum albumin: Dissolve 0.5 g bovine serum albumin in 50 mL M9 minimal medium (room temperature) and filter sterilize the resulting solution using a PES bottle top filter into an autoclaved 100 mL bottle. Store at 4
C. 3. Dye to stain cell membrane: PKH67 dye (Sigma), keep at 4
C until use. Then dilute 50 in Diluent C, at room temperature. 4. Diluent C (Sigma, included in the PKH67 kit). Store at 4
C, allow to warm to room temperature before the start of the staining. 5. M9 minimal medium without carbon source, at room temperature. 6. M9 minimal medium with fumarate added to a final concentration of 2 g/L, preheated to 37
C and pre-aerated.

2.3 Assessment of Antibiotic Tolerance with Flow Cytometry

1. Antibiotics, prepare concentrated stocks of (see Note 3): (a) 100 mg/mL ampicillin: Dissolve 107 mg ampicillin sodium salt in 1 mL water. Filter-sterilize using a PES syringe filter and store at 20
C. (b) 20 mg/mL tetracycline: Dissolve 108 mg tetracycline hydroxide in 5 mL water. Filter-sterilize using a PES syringe filter and aliquot stock in portions of 1 mL and store at 20
C. Protect against light. (c) 50 mg/mL kanamycin: Dissolve 120 mg kanamycin sulfate in 2 mL water, filter-sterilize using a PES syringe filter, and store at 20
C. (d) 35 mg/mL chloramphenicol: Dissolve 35 mg chloramphenicol in 1 mL 100% EtOH, filter-sterilize using a PES syringe filter and store at 20
C. (e) 0.4 mg/mL trimethoprim: Dissolve 40 mg trimethoprim in 100 mL water. Filter-sterilize using a PES syringe filter and aliquot in portions of 1 mL and store at 20
C.

Nutrient Shift to Generate Antibiotic-Tolerant Cells

45

(f) 1 mg/mL ofloxacin: Dissolve 10 mg ofloxacin in 10 mL water. Filter sterilize using a PES syringe filter and aliquot in portions of 1 mL and store at 20
C. (g) 1.25 mg/mL rifampicin: Dissolve 12.5 mg rifampicin in 10 mL water. Filter-sterilize using a PES syringe filter and aliquot in portions of 1 mL and store at 20
C. (h) 10 mg/mL carbonyl cyanide m-chlorophenylhydrazine (CCCP): Dissolve 20 mg CCCP in 2 mL pure methanol, filter-sterilize using a PES syringe filter and store at 20
C. 2. LB medium: Add tryptone (to a final concentration of 1% w/v), yeast extract (0.5% w/v) and sodium chloride (1% w/v) to a final volume of 500 mL water. Autoclave for 15 min at 121
C and sterilize using a 0.22 μm PES filter to remove debris, which disturbs the flow cytometric analyses (see Note 4). Store at room temperature.

3

Methods

3.1 Generation of Large Fractions of Tolerant Cells

To investigate tolerant cells at population level it is important to generate a cell population that consists almost entirely of dormant cells, even in the presence of a carbon source. In this section, we describe how to generate such a population by exposing the cells to a sudden nutrient shift. The procedure starts with generating an exponentially growing culture on glucose and continues with the steps to perform an abrupt switch to a different carbon source. In this section, fumarate is taken as the second carbon source because it generates the highest percentage of tolerant cells. Other gluconeogenic carbon sources can also be used for the switch [7]. A schematic overview of the method is shown in Fig. 1a–c. 1. Streak out an E. coli strain (e.g., BW25113) on LB-agar and incubate overnight at 37
C. Full-grown plates can be stored up to one month at 4
C sealed with parafilm. 2. Generate a culture that is fully exponentially growing on glucose, meaning that for several hours the cell number has been doubling at its maximal rate. Such a culture can be obtained by applying steps 3–5 (see Note 5): 3. (Day 1) inoculate a 100 mL flask containing 10 mL M9 minimal medium supplemented with glucose to a concentration of 5 g/L with a single colony from an agar plate at around 5 p.m. Grow overnight at 37
C and shaking at 300 rpm till around 9 a.m. the next morning. 4. (Day 2) to prepare the next preculture, first, determine the cell concentration in this overnight culture. For instance, dilute an

46

Silke R. Vedelaar et al.

Fig. 1 Schematic overview of the procedure for generating persisters and for testing for their antibiotic tolerance. (a) Prepare an exponentially growing culture. Add a colony to 50 mL glucose minimal medium and grow overnight. Dilute this culture and grow during the day, dilute once more, and grow overnight. Make sure the culture is in exponential phase when starting the staining. (b) Wash or stain the cells to remove all residual glucose and optional, to stain the cells for tracking them with flow cytometry. (c) Add the washed or stained cells to fumarate minimal medium and follow their growth at different time points using flow cytometric analysis. (d) To test antibiotic tolerance, add antibiotics after 2 h in fumarate minimal medium. Incubate for 2 h and transfer 500 μL to 50 mL LB medium. Follow regrowth for 4 h using flow cytometry

aliquot of the culture 100, transfer 200 μL to a 96-well plate (in case the used flow cytometer samples from 96-well plates) and measure the cell count (e.g., in 20 μL) with a flow

Nutrient Shift to Generate Antibiotic-Tolerant Cells

47

cytometer. Calculate the concentration in cells/mL, C, by using the following formula: C ¼ c  d  50 where c is the number of cells counted in 20 μL and d is the dilution factor. Then determine the volume of culture to be added to a fresh flask for having a target cell density of 3  108 cells/mL at 5 p.m. on the same day as follows: V inoc ¼

td t ln 2=μ

2



C v

where Vinoc is the volume that has to be added to the flask, td is the target density (in cells/mL) at the end of the day, μ is the growth rate of the exponentially growing population (in h1), t is the time until the end of the day (in h), and v is the culture volume (in mL). Add the calculated volume to a flask containing preheated (37
C) M9 minimal medium with 5 g/L glucose, and grow the culture at 37
C and shaking at 300 rpm. 5. (Day 2) to perform the actual nutrient shift the next morning (day 3) with a culture that is fully exponentially growing on glucose, perform another subculturing step the evening before (on day 2) to have 3  108 cells/mL at the desired time point in the morning of day 3. To calculate the volume that has to be added to a new flask, repeat step 4. Because of the huge dilution that has to be made for the inoculation of this overnight culture, it is recommended to inoculate several parallel flasks with dilutions such that one surely obtains at least one culture at the right cell density (see Note 6). Also, to prevent lag phase behavior, which can mess up the timing, it is recommended to use preheated and pre-aerated medium. 6. (Day 3) to start the nutrient shift, wait until the culture has reached OD600 of 0.3–0.8 before continuing with the nutrient shift. When the culture reaches the desired OD600, measure the cell concentration of the culture with the flow cytometer. The cell concentration should be around 2–5  108 cells/mL. With the knowledge of the cell concentration, calculate the volume of the culture that contains 1.5  109 cells. When staining is applied, skip the next steps and go immediately to Subheading 3.2. 7. Wash the cells to remove any residual glucose. Therefore: 8. Transfer the calculated volume to a 15 mL falcon tube and spin down the culture for 5 min at 3000  g; 4
C, and discard the supernatant. 9. Wash the pellet with 5 mL ice-cold M9 minimal medium without carbon source by carefully pipetting. The ice-cold medium is used to slow down the cells’ metabolism. Spin

48

Silke R. Vedelaar et al.

down, remove supernatant, and wash once more with the same volume of M9 minimal medium. 10. Spin down the culture for 5 min at 3000  g; 4
C, discard the supernatant and resuspend the pellet in 1 mL M9 minimal medium without carbon source. 11. Transfer the cells to a 500 mL flask containing 50 mL preheated M9 minimal medium with 2 g/L fumarate. The cell concentration in the new flask is important because the initial cell density influences the fraction of cells that enter the tolerant state after the nutrient shift (see Note 7). With E. coli BW25113, this procedure results in about 99.99% of cells entering the tolerant state [7, 8]. With other E. coli strains, this fraction can be different. 3.2 Identification and Quantification of Tolerant Cells Using Membrane Staining and Flow Cytometry

To determine the fraction of cells that enter the tolerant state, for example, after the nutrient shift, one can stain the membrane of the cells with a fluorescent dye, follow the fluorescence of the cell population after the nutrient shift over time and analyze the resulting single-cell fluorescence data with a computational script. The dye we propose (PKH67, Sigma) stains the membranes of the cells and equally distributes over the two daughter cells after cell division. Thus, the fluorescence intensity of a cell halves with every division. Knowing the initial fluorescence intensity of the population after staining and comparing this with the fluorescence intensity of a cell at a later time point, one can calculate how often a cell has divided [7]. Specifically, to estimate the fraction of cells that have not started to divide, a Matlab script is used that fits a mathematical model to the time course data (i.e., cell counts, fluorescence intensity distributions). The model assumes that the fluorescence intensity of a cell decreases by half with each cell division and that cells have some autofluorescence. Further, it assumes exponential growth of the growing cells from some time point after the nutrient shift onward. As the fluorescence intensities of individual cells are not identical, the model fits bimodal distributions to the fluorescence intensity distributions determined at the different time points and estimates the growth rates of the two populations of cells based on the total cell counts determined and fitted bimodal distributions. The model, its rational, and mathematics are explained in detail in our previous paper [7]. To determine the fraction of dormant cells that emerge after a sudden nutrient shift and to estimate the growth parameters (including, among others, the growth rates of both populations), membrane staining is applied in combination with the nutrient shift method to generate tolerant cells, meaning that the washing steps 7–11 in Subheading 3.1 are replaced by the following staining protocol. A schematic overview of this method is shown in Fig. 1a–c.

Nutrient Shift to Generate Antibiotic-Tolerant Cells

49

1. Generate exponentially growing cells on glucose according to steps 1–6 of Subheading 3.1. 2. Before the shift to fumarate, cells are stained with a fluorescent dye. Before the staining, take Diluent C from the PKH67 kit from the fridge and allow it to warm to room temperature. The dye should stay in the fridge. 3. Determine the volume of the glucose culture that contains 1.5  109 cells. Transfer the calculated volume to a 15 mL falcon tube. Spin down the culture for 5 min at 3000  g; 4
C (Eppendorf centrifuge). Discard the supernatant very carefully by pipetting (see Note 8). 4. During the centrifugation mentioned in the previous step, prepare a mix of 500 μL of room-temperature Diluent C with 10 μL of the dye solution (see Note 9). 5. When performing steps 5–9 act fast and respect the timings (see Note 10). Resuspend the cell pellet in 500 μL of solely Diluent C (room temperature) by carefully pipetting up and down. Make sure that the pellet is fully dissolved and that there are no droplets on the sides of the tube. 6. Add the prepared dye solution (10 μL of dye in 500 μL of Diluent C at room temperature) to the cell solution and mix briefly by vortexing. Incubate the cells for exactly 3 min at room temperature (see Note 11). 7. Immediately after 3 min add 4 mL of ice-cold M9 medium with 1% (w/v) BSA and mix briefly by vortexing (see Note 12). 8. Centrifuge the cells for 5 min at 3000  g; 4
C. Discard the supernatant by pipetting and resuspend the cells in 5 mL of ice-cold M9 without carbon source and centrifuge the cells again (5 min; 3000  g; 4
C). 9. Wash the cells once more in the same manner. Resuspend the cells in 1 mL M9 medium without carbon source at roomtemperature and transfer them to the preheated M9 minimal medium with 2 g/L fumarate (see Note 13) to yield an OD600 of approximately 0.6, which corresponds to a cell concentration of ~5  108 cells/mL. Check the quality of the staining using flow cytometry by measuring a proper dilution of the culture. Well-executed staining should have only one peak on the FL1 (533 nm) signal, and the cells should be 100-fold brighter than unstained cells. 10. To obtain data that can be used with the Matlab script, the measurements must be done at specific times. The first time point must be gathered as soon as possible after the switch. 11. The next data time point should be one in which the growing population is becoming visible as a separate peak in the FL1 histogram. This time point varies depending on the conditions and carbon sources used (see Note 14).

50

Silke R. Vedelaar et al.

12. After taking these first time points, it is advisable to take measurements every 30–60 min, depending on the growth rate of the growing population, until the growing population count is at least equal, or higher than the non-growing population count. 3.3 Guide on How to Use the Matlab Script

The Matlab script and exemplary input files are provided on GitHub (https://github.com/molecular-systems-biology). The following steps describe the data acquisition and data format requirements. 1. To assess tolerant cells, it is important to generate tolerant cells as described in Subheading 3.1 and to stain them as described in Subheading 3.2. The staining needs to be of an appropriate quality for the script to work optimally (see Note 15). 2. The data needs to be gathered over the growth of the culture according to steps 10–12 in Subheading 3.2 (see Note 16). Certain factors need to be considered when gathering the data, for it to be usable with the script. The input data needs to be provided as a CSV file containing the flow cytometry measurements for each cell at each time point, without a header (File 1). In the Matlab script, the following input needs to be provided. (a) The number identifying the column of the CSV file containing the relevant fluorescence intensity (FI) data (in case of our example data as provided at GitHub, the number is 3—variable SC). (b) The scaling factor. Different flow cytometers have different sensitivity and numerical output values. This factor is used to accommodate for these differences and make sure that the data from different machines can be used within the script capabilities (variable maxval_new_FC). (c) The times when the samples were taken (variable tt). (d) The absolute cell concentration in the culture at the respective time points (variable cc). (e) The number of cells (or rows) in each CSV data file (variable g). (f) A specification of which of the data files should be used for the fitting. Depending on the nutrient switch, sometimes it takes hours before the growing population is numerous enough to be detected. The first time point used for the fit should have this population visible. Some data points can be excluded, for example, if the measurement has failed or has been inaccurate for any scientifically justified reason (variables indx and cc_indx).

Nutrient Shift to Generate Antibiotic-Tolerant Cells

51

Table 1 Overview of the parameters needed to be set for Matlab script. Left column: overview of parameters that need to be set. Right column: instructions on how to estimate these values Parameter

How to determine

Non-growing population initial FI mean

From “Fluorescence Data” figure—see Fig. 2a

Non-growing population FI stdev

From “Fluorescence Data” figure—see Fig. 2a

Non-growing population cell concentration

The cell concentration after the nutrient switch is usually a good initial guess, except for nutrient switches in which the non-growing population is small. Upper and lower bounds can be very relaxed

Growth rate of non-growing population

From previous experiments

Growing population initial FI mean From “Fluorescence Data” figure—see Fig. 2a Growing population FI stdev

From “Fluorescence Data” figure—see Fig. 2a

Initial growing population cell concentration

A guesstimate based on the expected number of adapting cells for a particular carbon source switch. Upper and lower bounds can be very relaxed

Growth rate of the population that starts to grow normally

From previous experiments determining growth rate on particular carbon source

Unstained cell background autofluorescence

By analyzing data for unstained cells, or by checking the FI for very late samples, when cells have lost all their fluorescence

In the Matlab script, one needs to define the allowed ranges of the parameters to be estimated as well as initial parameter guesses. These values should be close to the true values. Table 1 provides an overview of the parameters that need to be set and instructions on how to estimate these values, for example, by visual inspection of the data or by using previous knowledge. It is advisable to run the script once on the data before setting the parameters. From the then generated “Fluorescence Data” figure (Fig. 2a), one can identify initial guesses (in matrix variable IG) and ranges for some of the parameters (in matrix variable bounds). 1. Two figures generated by the script are crucial at determining the quality of the fit, that is, the figure “Cellcount curve fit check” and the multipanel figure “Bi-gaussian fit for each time point.” An example of a good fit is shown in Fig. 2; two examples of bad fits are shown in the Supplementary Figs. 1 and 2. The files for these fits are provided with the Matlab script. The two respectively generated figures show the cell count and fluorescence intensity data along with the plotted model predictions for both the growing and non-growing populations, and the sum of these populations. The way the model prediction fits the data can be used to assess whether the parameters are estimated correctly.

52

Silke R. Vedelaar et al.

Fig. 2 Finding initial parameters and example of a good fit. (a) How to find the initial parameter guesses using the Fluorescence Data figure generated by the script. The model needs a good estimation input to generate proper estimations. Therefore, the mean fluorescence of the growing and the non-growing population needs to be estimated. To make a good estimation pick the average of the fluorescence in between the left arrows for growing cells and the average of the peak on the right for non-growing cells. (b) The cell count curve fit check. Empty disk—cell count at t ¼ 0, red disks—cell counts used for the model fit, red line—predicted total cell count, cyan line—cell count of non-growing cells, magenta line—cell count of growing cells. (c) The model fit to the fluorescence data at each time point. Blue line—experimental data, green line—the distribution corresponding to the growing population, red line—the distribution corresponding to the non-growing population, black line (not visible in this specific case)—the sum of the distributions pictured by red and green line

2. In the first bad fit (Supplementary Fig. 1), the parameter bounds are set wrong. The FI means for both populations are overestimated, and the bounds are outside of the correct values. Moreover, the nongrowing population cell concentration is underestimated. 3. From the plots (Supplementary Fig. 1a), certain problems can be deduced. (a) The red line on the left panel that describes the modeled cell counts does not appear to fit with the cell count points from measurements. This is caused by the underestimation of the non-growing population cell concentration.

Nutrient Shift to Generate Antibiotic-Tolerant Cells

53

(b) The green line and the red line on the right panel do not fit the bimodal distribution. This is caused, on top of cause from problem 1, by the overestimation of the FI means. 4. Moreover, we can see from the text output of the script that some of the parameters—the growth rate of the growing population in this case—reaches the lower limit of the bounds set, and thus it is not estimated correctly. However, we know what the growth rate of the used strain in given conditions should be and the fact that the estimate is equal to one of the boundaries is just a symptom of another issue, and not the issue itself. 5. If we fix the first problem and correct the nongrowing population cell concentration to a value that is close to the measured value, we obtain graphs shown in Supplementary Fig. 2. While the graph on the left looks like it is a good fit, the graph on the right has the similar problems as before, stemming from the overestimation of the FI means. Fixing these values as described above, we can obtain a good fit. 6. It is clear from the plots shown in Fig. 2b, c that the model closely fits the data and the data output can be trusted. 7. The parameters obtained from the fitting are output as text. The Matlab script used as an example generates the following data. alpha = 0.00047173 mu_growin = 0.43374 mu_nongrw = 0.0026477 sgma_grow = 53.0783 sgma_nong = 55.9828 I_0 = 3419.6016 EI = 20.0569 dye_degr = 0.015 BG cutoff = 1 BG magnit = 1 BG width = 1 CC weight = 1 Bigaus pt = 16.0833 16.8333 17.5833 18.3333 19.0833 19.8333 20.5833 21.3333 22.0833 22.8333 23.5833 24.3333 25.0833 25.8333 CC pts = 16.0833 16.8333 17.5833 18.3333 19.0833 19.8333 20.5833 21.3333 22.0833 22.8333 23.5833 24.3333 25.0833 25.8333

8. The most important obtained values are alpha (fraction of cells that entered dormancy after the nutrient shift), mu_growin (growth rate of the growing population), mu_nongrw (growth rate of the nongrowing population). All these and other parameters, except alpha, should be checked against the lower and

54

Silke R. Vedelaar et al.

upper bounds set in the script and if they are equal to them, the bounds should be relaxed until they do not limit the results anymore. 3.4 Assessment of Antibiotic Tolerance with Flow Cytometry

The ability to regrow after antibiotic treatment is an indication that a cell has been in the tolerant state during treatment. While typically such regrow experiments are done with plating assays (i.e., determination of colony-forming units), here we proposed a method to perform such regrow assessments with flow cytometry. These assays require the above mentioned membrane staining. Despite the staining procedure that needs to be carried out, this technique is less laborious and has less variability than plating assays [9]. Also, it was found that more cells were able to wake up after dormancy when liquid medium is used compared to regrowth on plates [10]. A schematic overview of this method is shown in Fig. 1d. 1. To assess the antibiotic tolerance of cells, it is important to generate tolerant cells as described in Subheading 3.1 and to stain them as described in Subheading 3.2. 2. After transfer to the fumarate medium, it takes a certain time for the non-adapting cells to develop full tolerance against antibiotics. It is recommended to let the cells adapt for at least 2 h after staining (see Note 17) before treating them with antibiotics. 3. After >2 h, add the antibiotics. We tested 100 μg/ml ampicillin, 20 μg/mL tetracycline, 100 μg/mL kanamycin, 140 μg/ ml chloramphenicol, 5 μg/mL trimethoprim, 100 μg/ml rifampicin, 5 μg/ml ofloxacin, and 50 μg/ml CCCP [8]. Concentrations were obtained from literature and from survival assays on cells growing on glucose (see Note 18). Because some cells undergo a reductive division after the switch to fumarate, the number of cells after 2 h is slightly higher than at time point zero. Therefore, it is required to measure a dilution of the culture directly after staining and right before antibiotics are added. 4. Incubate the cultures that now contain the antibiotics for 2 h whilst shaking at 300 rpm at 37
C. After the incubation time, measure again a dilution of the culture with the flow cytometer to determine the cell count. The cell count should not have increased after the addition of antibiotics. In some cases, the cell count can even be decreased, for example, when a bacteriolytic antibiotic is used (see Note 19). 5. To assess the fraction of cells being able to recover from antibiotic treatment, the antibiotics are diluted out. Pipet 500 μL of the culture into 50 mL preheated and filtered LB medium (see Note 14), mix well and transfer a sample of the non-diluted culture in the 96-well plate (from where the flow cytometer will

Nutrient Shift to Generate Antibiotic-Tolerant Cells

55

take the sample) and measure it straight away with the flow cytometer (see Note 20). The first sample taken from the LB culture is very important because it reflects all cells in their dormant state and it is used to calculate the fraction of cells that became dormant after the nutrient switch (see Note 21). 6. From now on, subject the LB cultures to flow cytometric analyses every 30 min for a period of 4 h. Cells have a short doubling time on LB and thus cell counts can quickly enter a range that is beyond the linear range of the flow cytometer. Take notice of the cell count in each sample and determine the appropriate dilution for the next sample (see Note 22). 7. The fraction of tolerant cells is calculated by subtracting the number of non-dividing cells (cells that do not exit the bottom right part of Fig. 3a) in each time point from the initial number of nondividing cells.

Fig. 3 Example of regrowth in LB medium after antibiotic treatment. (a) Exemplary flow cytometry graphs over time. When tolerant cells (stained with fluorescent dye) are transferred to LB, they first increase in size followed by a loss in fluorescence as a consequence of their divisions. Cells in Q1 are big and have lost their fluorescence. Cells in Q2 are big and have a high fluorescence intensity. Cells in Q3 are small and have lost their fluorescence. Cells in Q4 are small and are fluorescent. (b) Left, the formula of how the fraction of tolerant cells is calculated. Right, an example graph of treatment with two different antibiotics. The fraction of cells for each time point is calculated by 1 minus the number of cells in Q4 divided of the number of cells in Q4 on time point 0. Cells that are killed by antibiotics will not regrow in LB medium and therefore will not leave section Q4 in the flow cytometer graph. After 4 h, a steady state is reached and the fraction of cells on t ¼ 4 can therefore be used as the ultimate fraction of viable cells

56

Silke R. Vedelaar et al.

8. Samples measured directly after staining, after the addition of antibiotics, and right before the nutrient shift are solely used as quality controls for the experiment. The cells should show one single small peak directly after the switch to fumarate and the population should have significantly slowed-down growth before antibiotics are added. After the 2 h treatment with antibiotics the growth should have completely stagnated and the cell number can even drop in case a bacteriolytic antibiotic (e.g., ampicillin) is used. 9. For data analysis of the total fraction of tolerant cells, the cells are followed for 4 h after the switch to LB, and the loss of fluorescence as well as the size of the cells is used to determine the fraction of the population being able to escape dormancy after antibiotic treatment. It is required to set up a dot plot to gate the cells (FSC-H vs SSC-H) and a dot plot using cell size and fluorescence (FL1-A vs FSC-A). 10. To determine the fraction of cells escaping dormancy, the number of cells transferred to LB is taken at time point zero (T0). For each following time point, the number of nondividing cells is subtracted from the cells transferred at T0. It is not possible to use the number of escaping cells since they start dividing after the switch to LB (see Note 23) as shown in Fig. 3a. 11. To define the non-growing population, the quadrant function, as in Fig. 3a, needs to be adjusted manually for each sample. Especially in samples where a big part of the population just started to divide, it can be hard to distinguish the non-growing population from the growing population. It is advised to use the later time points to set the quadrant and use the same settings for the early division samples (see Note 24). When one is only interested in the final number of tolerant cells and not in the recovery dynamics, one can decide to not analyze these samples. An example of the recovery dynamics is shown in Fig. 3b. 12. The data from 4 h on LB is used to calculate the final fraction of tolerant cells. Take the number of nondividing cells and subtract it from the number of cells added to LB at T0. It is optional to take the average of the last two, or the last three measurements to calculate the final fraction of tolerant cells (only if the recovered fraction has reached steady state). Averages and standard deviations are calculated from three individual experiments and shown as box plots.

Nutrient Shift to Generate Antibiotic-Tolerant Cells

4

57

Notes 1. If antibiotics are required it is important that the antibiotic is always added after autoclaving the LB-agar and after the LB agar has cooled down to ~55
C, to avoid degradation of the antibiotic. It is recommended to prepare 1000 concentrated stocks of the required antibiotics, to store them at 20
C and thaw them just before pouring plates. 2. When using the glucose-to-fumarate rapid nutrient shift to generate large fractions of tolerant cells, it is important to not use cellulose acetate filters because these filters could release acetate into the medium [13], which E. coli can use as a carbon source. As E. coli can have different preferences on which gluconeogenic carbon source to use first, even a small concentration of acetate could affect the results. PES filters are equally priced and do not release any compound that can act as a substrate for E. coli. 3. The antibiotic concentrations used in the working solutions are based on literature research and on experiments where we tested the antibiotics on E. coli cells growing on fumarate. The concentration of the antibiotic stocks is based on the solubility of the antibiotic. For ampicillin and tetracycline, the solubility is very high so it was decided to make a 1000 concentrated stock. 4. When measuring and counting E. coli cells, the flow cytometer needs to be calibrated for small cells. Typically, the default settings are not suitable for E. coli cells. With the BD Accuri C6, the threshold for E. coli should be 8000 for the forward scatter (FSC-H) and 500 for the side scatter (SSC-H). Further, debris in the medium can disturb the measurement. Thus, it is required to only use filtered medium when measuring samples with the flow cytometer. For reliable cell counts with the Accuri C6 flow cytometer, the events measured by the flow cytometer should be between 10,000 and 100,000 cells in 20 μL of the measured sample. This level ensures a sufficiently high number of cells in comparison to machine noise, and is sufficiently low to not cause over-saturation of the instrument. For LB it is important to autoclave and filter sterilize the medium before use in the flow cytometer. LB contains, when only autoclaved, a lot of components that lead to a strong background signal in the flow cytometer. This interferes with the sample because the debris appears around the same size as your cells on the FSC-SSC dot plot and therefore it is required to use filtered LB medium to lower the background signal. 5. For the successful generation of almost 100% tolerant cells with a glucose-to-fumarate shift, the cells in the glucose culture must be in a fully exponentially growing state on glucose for

58

Silke R. Vedelaar et al.

at least 24 h before the nutrient shift. Because cells typically have a lag phase after inoculation from the LB plate, we recommend to start culturing cells at least 1 day before the actual nutrient shift to ensure the maximal growth rate on glucose. 6. To ensure that cells keep on growing exponentially also after the re-inoculation, it is crucial to pre-heat the medium before dilution of the cultures. Even then, typically short lag phases occur after diluting, and it is therefore recommended to not only inoculate the calculated number of cells but also to prepare a preculture starting with 2 the number of cells required. In so doing, at the desired time point in the morning one surely obtains a culture with the proper cell density. 7. Previous research has shown that the initial cell density after the nutrient shift influences the fraction of dormant cells after a nutrient shift [14]. The absolute number of adapting cells seems to be constant regardless of the initial cell density after a glucose-to-fumarate shift, resulting in a higher fraction of cells adapting when a flask is inoculated with a low cell density [13]. Furthermore, note that with different E. coli wild-type strains, different fractions of dormant cells can emerge. 8. Do not lose any cells when applying the staining, because the dye-to-cell number ratio is very important. Losing cells will increase the dye-to-cell ratio causing cells to be stained too intensely, which in turn can lead to cell death. A suboptimal dye-to-cell number ratio can also affect the viability of the cells, resulting in fewer cells being able to recover after antibiotic treatment. 9. Preparing this mixture too soon will cause the dye to clump and staining intensity will be suboptimal. The solution can be prepared during the first centrifugation step, although the Diluent C must be brought to room temperature earlier. When multiple samples are stained in parallel, always change the pipette tip to prevent water transfer to the vial containing the dye stock. The dye-to-cell number ratio is very important for proper staining. The ratio is good when the cells clump a bit during staining, as indicated by a slightly cloudy solution. 10. Leaving droplets on the side of the tube will create a fraction of unstained cells because droplets containing cells do not get in touch with the dye. This can be seen as a separate unstained fraction and will influence results or even make them unusable. 11. Longer incubation will cause the cells to die, whereas shorter incubation will cause the cells to be stained suboptimally. If you stain cells of multiple different samples in parallel, add the dye in 20–30 s intervals, such that the incubation time can be strictly adhered to.

Nutrient Shift to Generate Antibiotic-Tolerant Cells

59

12. BSA blocks the remaining dye molecules, thereby preventing the cells from being killed. It is recommended to use a timer and make sure you have the M9 medium with BSA in your pipette ready so that you only have to release it from the pipette into the tube containing the cells when the 3 min incubation time is over. 13. It is recommended to prepare the flasks with preheated and pre-aerated M9 medium with 2 g/L fumarate before harvesting the cells. 14. The higher the fraction of adapting cells, the sooner the first usable data point can be taken. In the case of a glucose-tofumarate switch, 15–16 h after the switch is usually a good starting point, as the growing population starts to be large enough to be detectable, while it still has a fluorescence intensity that is above the autofluorescence. Capturing time points in which the growing population still has some fluorescence above the cellular autofluorescence is crucial for the script to estimate the fraction of growing cells accurately. 15. To obtain a good fit of the data to the model and thus reliable parameter estimates, the data needs to fulfill certain criteria. First, the fluorescence intensity of stained cells should be at least two orders of magnitude higher than the background fluorescence of unstained cells. This, in turn, means that the fluorescence intensity decrease can be tracked over 5–6 divisions, which is enough for most carbon source switches. Moreover, the fluorescence intensity data of stained cells should form a single peak in the histogram. The narrower the peak, the better, as the growing and nongrowing populations will be better separated from each other in the obtained data. While in our experiments we had the best results using the green fluorescent dye, we have also used a red fluorescent dye (PKH26), which has giving us data of lesser quality. In principle, any product that utilizes a fluorophore linked to an aliphatic chain that intercalates in the cell membrane could be utilized to generate the data for the script, and there are many alternative products available on the market that have different excitation and emission properties that might be best suitable for the equipment used. However, due to different protocols and properties of these dyes, appropriate controls would have to be made to exclude the effect of the staining procedure on the obtained results. 16. The fluorescence intensity data from the flow cytometer needs to be in log10 space. Depending on the cytometer used, a transformation needs to be done on the data (as we do in case of the Accuri C6 cytometer) or is already done in the cytometer or its software. Moreover, depending on the flow cytometer, the numerical range of values can be different and this needs to be addressed by setting the scaling factor.

60

Silke R. Vedelaar et al.

17. After transfer to the fumarate medium, several cells will undergo a reductive cell division within the first hour. Experiments have shown that after 2 h on fumarate, all cells of the population become tolerant against ampicillin [8]. Adding antibiotics too early might result in fewer cells surviving treatment, whereas incubating longer than 2 h does not increase the size of the tolerant population. 18. For each antibiotic used in our experiments, we checked in the literature which concentrations were used in E. coli inhibition experiments. We used this information as a starting point to explore which concentration of each antibiotic kills growing cells. Before the tolerance experiments were done we carried out identical experiments with glucose-grown cells to determine the concentration of antibiotics needed to kill a growing population. 19. When switching the cells to fumarate, 0.01% of the population will adapt and start proliferating [7]. Those cells will be sensitive to antibiotic treatment. Especially when the tolerant cells are kept for longer periods (~24 h after the nutrient shift), a significant population of growing cells is visible which are sensitive to antibiotics. When a bacteriolytic antibiotic such as ampicillin is used, the treatment will cause these cells to lyse. Therefore, the cell count after treatment can be lower than before treatment. 20. The sample should not be diluted because there are now 100 times less cells in the culture than in the flask where the cells were incubated with the antibiotics. 21. To check the cells’ ability to survive antibiotic treatment they are transferred to LB medium. No washing steps are applied to protect the cells from extra stress. By only transferring 500 μL in 50 mL the antibiotics are diluted 100-fold, enough to nullify their inhibiting effect. 22. Bacteria grow fast on LB medium, μ ¼ 1.9 h1 [15]. Therefore, to generate reliable data points it is essential to take regular measurements. Dead cells do not proliferate, meaning that their number remains the same and they do not lose their fluorescence. They are visible as a stained cloud of small cells in the FL1-A versus forward scatter (FSC-A) plot. However, when the descendants of the small growing population keep increasing their number, they outgrow the linear range of the flow cytometer and a bigger dilution must be made to measure the sample. This has the risk to lose the visibility of the population of dead cells because they are excessively diluted and their cell count is not reliable. In particular, when only a small fraction of cells are not dividing, a small dilution introduces a big measurement error by diluting out the number of non-growing cells.

Nutrient Shift to Generate Antibiotic-Tolerant Cells

61

23. Bacteria escaping the dormant state do this by resuming growth. Since it is impossible to determine the individual growth rate for each cell, there is no way to use the growing population for calculating the fraction of cells able to escape dormancy. However, since we know the original number of cells transferred to LB and we can distinguish growing from nondividing cells, we can use the number of nondividing cells and the number of cells at T0 to calculate the fraction of the population which has survived antibiotic treatment. 24. When cells start waking up from their dormant state, they will first get bigger, followed by the loss of fluorescence. This can be seen in a dot plot as in Fig. 3a (FL1-A vs FSC-A). However, in some cases, the cells may wake up a bit slowly, and in the earlier time points of the recovery experiment, the population of cells with increasing size might overlap with the nondividing population. In that case, it is advised to determine the nondividing population at a later time point and use those settings in the more indefinite sample.

Acknowledgements This work was supported by the Netherlands Organisation for Scientific Research (NWO) through a VIDI grant to MH [project number 864.11.001]. References 1. Balaban NQ, Helaine S, Lewis K et al (2019) Definitions and guidelines for research on antibiotic persistence. Nat Rev Microbiol 17:441–448 2. Lewis K (2010) Persister cells. Annu Rev Microbiol 64:357–372 3. Feng J, Kessler DA, Ben-Jacob E, Levine H (2014) Growth feedback as a basis for persister bistability. Proc Natl Acad Sci U S A 111:544–549 4. Moyed HS, Bertrand KP (1983) hipA, a newly recognized gene of Escherichia coli K-12 that affects frequency of persistence after inhibition of murein synthesis. J Bacteriol 155:768–775 5. Nguyen D, Joshi-Datar A, Lepine F et al (2011) Active starvation responses mediate antibiotic tolerance in biofilms and nutrientlimited bacteria. Science 334:982–986 6. Amato S, Brynildsen M (2015) Persister heterogeneity arising from a single metabolic stress. Curr Biol 25:2090–2098

7. Kotte O, Volkmer B, Radzikowski JL, Heinemann M (2014) Phenotypic bistability in Escherichia coli’s central carbon metabolism. Mol Syst Biol 10:736 8. Radzikowski JL, Vedelaar S, Siegel D et al (2016) Bacterial persistence is an active σS stress response to metabolic flux limitation. Mol Syst Biol 12:882 9. Jett BD, Hatter KL, Huycke MM, Gilmore MS (1997) Simplified agar plate method for quantifying viable bacteria. BioTechniques 23:648–650 10. Wong FH, Cai Y, Leck H et al (2020) Determining the development of persisters in extensively drug-resistant Acinetobacter baumannii upon exposure to polymyxin B-based antibiotic combinations using flow cytometry. Antimicrob Agents Chemother 64:e01712–e01719 11. Rousselle C, Barbier M, Comte V et al (2001) Innocuousness and intracellular distribution of PKH67: a fluorescent probe for cell

62

Silke R. Vedelaar et al.

proliferation assessment. In Vitro Cell Dev Biol Anim 37:646–655 ˜ ers A, Liske E, Tenson T (2020) Dividing 12. Jo subpopulation of Escherichia coli in stationary phase. Res Microbiol 71:153–157 13. Vos KD, Burris FO, Riley RL (1966) Kinetic study of the hydrolysis of cellulose acetate in the pH range of 2–10. J Appl Polym Sci 10:825–832

14. Radzikowski J (2012) Systems biology of bacterial persistence, a metabolism-driven strategy for survival. University of Groningen, Dissertation, Groningen 15. Schmidt A, Kochanowski K, Vedelaar S et al (2016) The quantitative and conditiondependent Escherichia coli proteome. Nat Biotechnol 34:104–110

Chapter 4 Enrichment of Persister Cells Through B-Lactam-Induced Filamentation and Size Separation Etthel Windels , Bram Van den Bergh , and Jan Michiels Abstract Analyzing persisters at the single-cell level is crucial to properly define their phenotypic traits. However, single-cell analyses are challenging due to the rare and temporary nature of persister cells, thus requiring their rapid and efficient enrichment in a culture. Existing methods to isolate persisters from a bacterial population show important shortcomings, including contamination with susceptible cells and/or cell debris, which complicate subsequent microscopic analyses. We here describe a protocol to enrich persisters in a culture using β-lactam-induced filamentation followed by size separation. This protocol minimizes the amount of cell debris in the final sample, facilitating single-cell studies of persister cells. Key words Persister, Enrichment, Single-cell

1

Introduction As bacterial persistence involves phenotypic heterogeneity within a population, interrogation of the persister physiology requires the selective enrichment of persister cells in a population. Many standard microbiological and molecular techniques run into limitations at this point, as persisters are often present at very low frequencies and existing fluorescent markers do not reliably distinguish persisters from susceptible cells, resulting in samples that are highly contaminated with susceptible cells [1–3]. Persisters can be enriched in a culture through prolonged antibiotic treatment that kills susceptible cells, as antibiotic tolerance is the only feature that is currently known to characterize all persister cells [4, 5]. However, the resulting sample is dominated by dead cell debris and is therefore inappropriate for microscopic studies. Using an antibiotic that lyses susceptible cells, followed by sedimentation of intact cells, only slightly reduces contamination by cell debris [6]. In addition to low final persister densities, the above-described methods involve prolonged antibiotic exposure, which potentially affects

Natalie Verstraeten and Jan Michiels (eds.), Bacterial Persistence: Methods and Protocols, Methods in Molecular Biology, vol. 2357, https://doi.org/10.1007/978-1-0716-1621-5_4, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2021

63

64

Etthel Windels et al.

Fig. 1 Microscopy images of a culture treated with cephalexin. During treatment with 50μg/ml cephalexin, susceptible exponential-phase cells filament severely, followed by cell lysis (scale bar: 20μm). Performing size separation at maximal cell length, but before lysis is initiated, allows for efficient enrichment of antibiotictolerant cells (Reproduced from [10])

persister formation [7–9]. We here propose a persister enrichment method that overcomes these technical hurdles. This method exploits the unique killing properties of cephalexin, a β-lactam antibiotic that first induces strong filamentation before lysis is initiated (Fig. 1). Enriching antibiotic-tolerant persisters by size separation right before lysis of susceptible cells occurs, limits the antibiotic exposure time and results in a sample containing a high density of persister cells and very little cell debris, making it particularly useful for subsequent (single-cell) microscopic studies [10].

2

Materials Prepare all media using deionized water. Store at room temperature, unless indicated otherwise. 1. Antibiotic stock solution (1000): Weigh 50 mg of cephalexin powder stored at 4  C and add sterile ultrapure water (resistivity of 18.2 MΩ/cm at 25  C) to a final volume of 1 ml (see Note 1). Filter-sterilize (0.22μm) and store aliquots immediately at 20  C. 2. Mueller Hinton Broth (MHB): Follow the instructions of the supplier of premixed MHB powder (see Note 2). 3. 10 mM MgSO4 solution: Weigh 2.46 g of MgSO4∙7H2O and add water to a final volume of 1 L before autoclaving. 4. Sterile microcentrifuge tubes (1.5 ml). 5. Sterile plastic tubes (50 ml), suitable for centrifugation. 6. Sterile glass test tubes and Erlenmeyer flasks (250 ml). 7. Sterile Erlenmeyer filtration flasks (500 ml) with side arm.

Enrichment of Persister

65

8. Bu¨chner funnel, including a supporting adapter cone that fits the filtration flasks. 9. Polyvinylidene fluoride membrane filters with a diameter corresponding to the Bu¨chner funnel diameter and a pore size of 5μm (see Note 3). 10. Vacuum pump with connection tubing that fits the filtration flasks. 11. Microcentrifuge capable of 4000  g. 12. Centrifuge for 50 ml tubes capable of 4000  g. 13. Incubator at 37  C, capable of orbital shaking at 200 rpm.

3

Methods

3.1 Installation of the Vacuum Filtration System

Work at room temperature and under sterile conditions. Incubation is carried out at 37  C, shaking at 200 rpm. A schematic overview of the method is provided in Fig. 2 (see Note 4). 1. Connect the vacuum pump to the side arm of a sterile filtration flask. 2. Place the Bu¨chner funnel with adapter onto the filtration flask and make sure the flask is sealed airtight. 3. Place a membrane filter wetted with sterile growth medium in the Bu¨chner funnel.

3.2 Persister Enrichment Protocol

1. Inoculate the strain under investigation in a test tube containing 5 ml MHB and incubate for 20 h.

Fig. 2 Schematic overview of persister enrichment protocol. A bacterial culture of 106 CFU/ml is treated with 50μg/ml cephalexin for 60 min. During treatment, susceptible cells strongly filament but do not yet lyse, while persisters are not affected by the antibiotic. The treated culture is vacuum filtered (pore size of 5μm) to separate short, antibiotic-tolerant persisters from filamented, susceptible cells. After filtration, the culture is centrifuged to wash away the antibiotic and to increase the density of the resulting sample (Reproduced from [10])

66

Etthel Windels et al.

2. Dilute the overnight culture 1:5000 in 100 ml MHB and incubate for 90 min (see Note 5). When working with Escherichia coli, a cell density of 106 CFU/ml should be attained at this point (see Notes 6 and 7). 3. Treat the culture with 50μg/ml cephalexin for 60 min (see Notes 8–10). During treatment, antibiotic-susceptible cells should strongly filament but not yet lyse (Fig. 1). 4. Pour the culture in the Bu¨chner funnel and perform vacuum filtration (see Note 11). 5. Replace the filtration flask containing the filtered culture with an empty, sterile flask. Replace the used membrane filter with a fresh filter. 6. Pour the filtered culture in the Bu¨chner funnel and perform vacuum filtration. 7. Collect the filtrate in sterile 50 ml tubes. 8. Centrifuge the tubes at 4000  g for 5 min. Pour off the supernatant. 9. Resuspend the pellet in the remaining volume. Transfer this volume to a microcentrifuge tube. Pooling cells from different tubes into one microcentrifuge tube increases the cell density in the final sample. 10. Centrifuge the microcentrifuge tubes at 4000  g. Pipette off the supernatant and resuspend in 500μl 10 mM MgSO4 by vortexing. 11. Centrifuge the microcentrifuge tubes again at 4000  g. Pipette off the supernatant and resuspend in a small volume of 10 mM MgSO4 or MHB, depending on the subsequent experiment. For microscopy experiments, resuspend the pellet in a few μl of 10 mM MgSO4 or MHB by pipetting up and down. A representative example of a resulting sample is shown in Fig. 3.

4

Notes 1. Add drops of 1 M NaOH to increase the solubility of cephalexin in deionized water. 2. Using premixed MHB formulas may enhance reproducibility. Alternatively, MHB can be prepared using 2 g/L beef extract powder, 17.5 g/L acid digest of casein, and 1.5 g/L starch. Store the autoclaved medium at room temperature and away from light. Do not store the medium for longer than a week after autoclaving. Never autoclave the medium twice.

Enrichment of Persister

67

Fig. 3 Microscopy image of a sample after cephalexin treatment and vacuum filtration. The final sample contains antibiotic-tolerant cells that did not filament in response to cephalexin treatment (scale bar: 20 μm). As filtration is performed before susceptible cells start lysing, the sample contains a very limited amount of cell debris (Reproduced from [10])

3. Filters with a pore size of 5μm are convenient for E. coli, as normal E. coli cells have an average cell length of 2μm while cephalexin-treated cells elongate to 10–100μm. Adjust the pore size according to the organism under study and the filamentation effect of the used antibiotic. 4. The protocol described in this chapter is validated for E. coli K-12 MG1655 in MHB medium with cephalexin treatment. Optimization of the protocol is advisable when using other strains, media or antibiotics (see Notes 6–8 and 10). 5. Depending on the desired number of persisters in the final sample, a larger culture volume can be used. 6. When using other strains or species, take into account that the efficacy of cephalexin is highly dependent on the cell density at the onset of treatment. The overnight culture should be sufficiently diluted to ensure a low cell density when treatment is initiated (106 CFU/ml, although this depends on the treatment concentration of cephalexin). For E. coli K-12 MG1655, stronger dilution results in less persisters in the final sample, while weaker dilution results in less efficient cephalexininduced elongation. 7. When using other strains or species, take into account that the incubation time after dilution and before treatment should be sufficiently long in order for susceptible cells to escape the lag

68

Etthel Windels et al.

phase. However, incubating the culture for too long results in a too high cell density at the onset of treatment (see Note 6). An optimal combination of the dilution factor and the incubation time before treatment should be found in order to achieve an efficient cephalexin treatment. 8. Cephalexin can be replaced by another antibiotic with similar, filamentation-inducing effects (e.g., ciprofloxacin or aztreonam). For an efficient size separation, it is important that the size of filamented cells sufficiently differs from nonfilamented cells. 9. For E. coli K-12 MG1655, a cephalexin treatment concentration of 50μg/ml corresponds to 6.25 the minimum inhibitory concentration. 10. When using another strain or antibiotic, the treatment concentration and treatment duration should be optimized. During treatment, susceptible cells should filament considerably without showing cell lysis. A treatment that is too short results in many susceptible cells passing the filter, while a treatment that is too long results in much debris from lysed cells passing the filter and contaminating the sample. 11. When using a culture volume larger than 100 ml (see Note 5), the filter might get clogged and filtration might be too slow. To prevent this, divide the culture into smaller volumes and replace the filter in between filtration of these volumes. This is not necessary during the second filtration step, as most elongated cells should have been removed from the culture at this point.

Acknowledgments E.M.W. and B.V.D.B. are Research Foundation Flanders (FWO)fellows. This research was funded by the KU Leuven Research Council (PF/10/010; C16/17/006), FWO (G0B0420N; G055517N), and the Flemish Institute for Biotechnology (VIB). References 1. Shah D, Zhang Z, Khodursky AB et al (2006) Persisters: a distinct physiological state of E. coli. BMC Microbiol 6:53 ˜ers A, Luidalepp H et al (2008) 2. Roostalu J, Jo Cell division in Escherichia coli cultures monitored at single cell resolution. BMC Microbiol 8:68 3. Sugimoto S, Arita-Morioka KI, Mizunoe Y et al (2015) Thioflavin T as a fluorescence probe for monitoring RNA metabolism at molecular and cellular levels. Nucleic Acids Res 43:e92

4. Keren I, Kaldalu N, Spoering A et al (2004) Persister cells and tolerance to antimicrobials. FEMS Microbiol Lett 230:13–18 5. Balaban NQ, Merrin J, Chait R et al (2004) Bacterial persistence as a phenotypic switch. Science 305:1622–1625 6. Keren I, Shah D, Spoering A et al (2004) Specialized persister cells and the mechanism of multidrug tolerance in Escherichia coli. J Bacteriol 186:8172–8180

Enrichment of Persister 7. Johnson PJT, Levin BR (2013) Pharmacodynamics, population dynamics, and the evolution of persistence in Staphylococcus aureus. PLoS Genet 9:e1003123 8. Kwan BW, Valenta JA, Benedik MJ, Wood TK (2013) Arrested protein synthesis increases persister-like cell formation. Antimicrob Agents Chemother 57:1468–1473

69

9. Do¨rr T, Vulic´ M, Lewis K (2010) Ciprofloxacin causes persister formation by inducing the TisB toxin in Escherichia coli. PLoS Biol 8: e1000317 10. Windels EM, Ben Meriem Z, Zahir T et al (2019) Enrichment of persisters enabled by ß-lactam-induced filamentation reveals their single-cell awakening characteristics. Commun Biol 2:426

Chapter 5 Methods for Enrichment of Bacterial Persister Populations for Phenotypic Screens and Genomic Studies Samantha Adikari , Elizabeth Hong-Geller , and Sofiya Micheva-Viteva Abstract Transient phenotypic adaptations in bacteria that enable survival at bactericidal antibiotic concentrations give rise to bacterial persistence. Naturally, the abundance of persister cells is very low (about 1 in 105 cells) in actively growing bacterial populations. Therefore, in order to study bacterial persistence mechanisms for therapeutics development, persister cells need to be enriched from a larger culture. Here, we describe three enrichment methods for obtaining Burkholderia thailandensis persisters: (1) flow sorting for persisters from exponentially dividing cultures by fluorescent staining of bacterial cells with a translational membrane depolarization-specific DiBAC4(3) dye, (2) antibiotic lysis of nonpersisters, and (3) culture aging to induce persister survival. We also describe herein the lysis of persister cells obtained by all three methods for downstream bacterial RNA extraction and transcriptomics analysis. Keywords Bacterial persistence, Persister enrichment, Antibiotic tolerance, Stress response, Nutrient limitation, Culture aging

1

Introduction Bacteria employ diverse mechanisms to overcome environmental stress conditions, including nutrient limitation and exposure to toxic metabolites. Phenotypic adaptations that increase bacterial survival under adverse conditions can also produce a small sub-population of transiently antibiotic tolerant cells. Current understanding of bacterial defense mechanisms causing antibiotic tolerance and persistence has been detailed in several review articles [1–3]. Various persister enrichment methods have been applied to study gene expression profiles responsible for the bacterial persistence phenotype. Bacteria exhibit high antibiotic tolerance in their stationary growth phase. Hence, aging of bacterial cultures is a simple means of naturally enriching persisters in a population

Natalie Verstraeten and Jan Michiels (eds.), Bacterial Persistence: Methods and Protocols, Methods in Molecular Biology, vol. 2357, https://doi.org/10.1007/978-1-0716-1621-5_5, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2021

71

72

Samantha Adikari et al.

[4, 5]. Other methods applied to enrich bacterial persisters include antibiotic lysis of susceptible bacteria [6, 7], and flow cytometrybased separation of genetically modified bacterial populations expressing dormancy markers [8]. It is notable that different stress conditions utilized to induce persistence in an experimental setting can unduly bias the outcome of phenotypic or gene activation assays. Thus, to obtain data representative of universal mechanisms that govern bacterial persistence, it is prudent to compare physiological and/or transcriptomic changes in response to a variety of different stressors that enrich the persister populations. In a recently published comparative transcriptomics study [9], we demonstrated that Burkholderia thailandensis (B.th.) persister populations display gene expression profiles that are specific to the enrichment method used. Nevertheless, a small subset of functional and regulatory gene pathways are activated universally in persister populations independently enriched via two of three or all three methods. Finding core genes regulating the persister metabolic program will greatly improve the design of clinically relevant therapies to treat persistent bacterial infections. Here we describe three distinct methods that we have previously applied to discover genes universally activated in B.th. persister populations: (1) flow sorting for persister cells exhibiting low proton motive force (PMF) in exponentially dividing populations; (2) lysing nonpersister cells with a beta-lactam antibiotic; and (3) inducing persister cell survival by exposing cultures to temperature stress and nutrient limitation. Although the culture and enrichment methods described herein have been applied to Burkholderia spp., which enter persistence at a higher rate than the majority of studied bacterial species [10–12], these methods could be utilized to enrich persister populations in other Gram-negative pathogens for various applications.

2

Materials Prepare and store all materials and reagents at ambient room temperature unless otherwise specified. Refer to the MSDS of each chemical and commercial reagent for details on hazard characteristics, and follow waste disposal regulations accordingly. Note that sodium azide is not added to any preparations made in-house. 1. LB agar plates: Add 1 L of ultrapure water (prepared by purifying deionized water to attain a sensitivity of 18 MΩ cm at 25
C) and 40 g of LB agar powder to an autoclave-safe glass bottle, close the lid, and shake vigorously to dissolve the powder. Autoclave for 20 min at 121
C and 18 psi. Transfer the bottle to a Class II biosafety cabinet (BSC), and pour hot agar into sterile petri dishes with lids (10 cm2 rounds and 6  6

Enrichment of Bacterial Persister Populations

73

grid-marked squares), about 25 mL per plate. Allow agar plates to cool without lids for 1 h (see Note 1). Once agar plates have solidified and dried, replace the lids, and stack agar plates (lid-side-down) inside a plastic sleeve. Store at 4
C for up to 1 month. Open only inside a Class II BSC. 2. LB Broth: Add 500 mL of ultrapure water (prepared by purifying deionized water to attain a sensitivity of 18 MΩ cm at 25
C) and 12.5 g of LB Broth powder to a glass bottle, close the lid, and shake vigorously to dissolve the powder. Transfer the bottle to a Class II BSC, and filter LB Broth through a vacuum-assisted 0.2μm PES membrane filter into a sterile receiver bottle. Store at ambient room temperature for up to 1 month. Open only inside a Class II BSC. 3. Burkholderia thailandensis E264 (BEI Resources) in 50% (v/v) glycerol stock, stored at 80
C. 4. Liquid chromatography-mass spectrometry (LC-MS) grade dimethyl sulfoxide (DMSO). 5. 10 mg/mL working stock of meropenem: Dissolve 50 mg of meropenem powder in 5 mL of DMSO. Make single-use aliquots (50μL each) in sterile 200μL tubes, and freeze immediately. Store at 80
C for up to 12 months (see Note 2). 6. 1 M/1 N NaOH (40 mg/mL). 7. 10 mg/mL working stock of ofloxacin: Dissolve 0.1 g of ofloxacin powder in 10 mL of 1 M NaOH. Make single-use aliquots (50μL each) in sterile 200μL tubes, and freeze immediately. Store at 80
C for up to 12 months (see Note 2). 8. DNase/RNase-free distilled water. 9. 100μg/mL working stock of bis(1,3-dibutylbarbituric acid) trimethine oxonol (DiBAC4(3)): Dilute the 1 mg/mL commercial reagent in a 1:10 ratio in DNase/RNase-free distilled water. Store at 20
C for up to 1 month. 10. 1 phosphate buffered saline (PBS) at pH 7.2, sterile, without Ca2+ or Mg2+ ions. 11. BacTiter-Glo Microbial Cell Viability Assay (Promega). 12. 5 mg/mL working stock of lysozyme: Dissolve 10 mg of lysozyme powder in 2 mL of 0.1 M KH2PO4. Make singleuse aliquots (20μL each) in sterile 200μL tubes, and freeze immediately. Store at 20
C for up to 12 months. 13. QIAzol Lysis Reagent (Qiagen) (see Note 3). 14. miRNeasy Mini Kit (Qiagen). 15. Chloroform.

74

3

Samantha Adikari et al.

Methods Perform all procedures at ambient room temperature unless otherwise specified. For all procedures prior to bacterial cell lysis, perform all steps inside a Class II BSC to ensure the sterility of materials and samples. For two of the three enrichment protocols described herein (reduced PMF and antibiotic lysis of nonpersisters), we isolated persister cells from exponentially dividing B.th. cultures. These protocols could be similarly applied to isolate persisters from stationary-phase liquid cultures, although the minimal bactericidal antibiotic concentration and the duration of antibiotic exposure are both expected to increase, and therefore would need to be empirically determined for those cultures.

3.1 Preparation of Exponentially Dividing Bacterial Cultures

1. Prepare a new colony stock plate of B.th. E264. Using a sterile pipette tip, streak B.th. E264 from the glycerol stock tube onto an LB agar plate. Incubate the plate for 24 h at 37
C. Store at ambient room temperature for up to 4 days (see Note 4). 2. Prepare a new stationary-phase liquid culture of B.th. E264. Using a sterile pipette tip, scrape a single colony of B.th. E264 from the colony stock plate, and resuspend into 5 mL of LB Broth in a 50 mL tube. Incubate the tube for 18 h at 37
C with agitation at 220 rpm. Store at ambient room temperature for up to 8 h (see Note 5). 3. Prepare a new mid-log phase culture of B.th. E264. Dilute the stationary-phase liquid culture (made the previous day) in a 1:500 ratio into fresh LB Broth in a 50 mL tube or 250 mL conical culture flask, depending on the culture volume required. Incubate tube/flask for 4–5 h at 37
C with agitation at 220 rpm. Use immediately for the relevant enrichment protocol when optical density at 600 nm (OD600), as measured by a spectrophotometer, reaches 0.20–0.30 (see Note 6).

3.2 Determination of Minimal Bactericidal Antibiotic Concentration

Antibiotic treatments were performed using meropenem, a betalactam antibiotic that inhibits cell wall synthesis, or ofloxacin, a quinolone antibiotic that inhibits DNA replication. The minimum inhibitory concentration (MIC) and minimum bactericidal concentration (MBC) of each new batch of antibiotics was determined prior to performing persister enrichment protocols. The MIC was calculated as the lowest antibiotic concentration at which the bacterial culture growth did not exceed the initial inoculation cell counts after 24 h of antibiotic exposure with aeration. The MBC was calculated as the antibiotic concentration at which a 50% loss of population viability was achieved.

Enrichment of Bacterial Persister Populations

75

1. Prepare 20 mL of a mid-log phase culture of B.th. E264 in a 50 mL Falcon tube (as described in Subheading 3.1, steps 1–3). 2. Thaw a single-use aliquot of antibiotic to test. 3. Label 5 1.5 mL sterile tubes: 10, 7.5, 5, 2.5, and 1 mg/mL. 4. Prepare at least five different antibiotic stock concentrations: 10, 7.5, 5, 2.5, and 1 mg/mL. Use solvent that is optimal for each antibiotic type; for example, meropenem stock concentrations are made in DMSO, while ofloxacin stock concentrations are prepared in 1 M NaOH. 5. Mix the mid-log phase culture by inverting the 50 mL tube, then remove 1 mL of culture to measure the OD600. When OD600 ¼ 0.20–0.30, proceed to the next step. 6. Set aside 1 mL of mid-log phase culture to determine viable bacterial colony-forming units per mL culture (CFU/mL) corresponding to the bacterial density prior to antibiotic treatment (as described in Subheading 3.2, steps 13–23). 7. Label 6 15 mL Falcon tubes: 10, 7.5, 5, 25, and 1μg/mL, untreated. 8. Mix the mid-log phase culture by inverting the 50 mL tube, then transfer 2 mL of culture to each labelled 15 mL tube. 9. Add 2μL of the corresponding antibiotic dilution to each tube (e.g., 2μL of 1 mg/mL stock into 1μg/mL tube). 10. Close the lids tightly and vortex the 15 mL tubes for 5 s to mix cultures thoroughly. 11. Incubate the 2 mL cultures for 24 h at 37
C with agitation at 220 rpm. 12. Immediately perform cell survival counts (as described in Subheading 3.2, steps 13–23), first for the untreated mid-log phase culture (Subheading 3.2, step 6), and then for the treated cultures (Subheading 3.2, step 11) after 24 h of agitation. 13. Label a 6  6 grid-marked square LB agar plate with dilutions from 101 to 106. 14. Transfer 200μL of bacterial culture from each treatment condition into the first row of a 96-well sterile U-bottom microtiter plate with lid. 15. Using a multichannel pipette, perform serial dilutions from 101 to 106 across the 96-well plate. Add 180μL of fresh LB Broth per well, then transfer 20μL of culture stepwise across the plate to achieve at ten-fold dilution in 200μL total, mixing thoroughly and changing pipette tips after each transfer.

76

Samantha Adikari et al.

16. Spot 10μL from each dilution (from 101 to 106) for every sample onto the labelled square agar plate, changing the pipette tip after each spot. 17. Allow spots on the square agar plate to dry for 10 min undisturbed, with the plate lid on. 18. Wrap the 96-well plate (used for serial dilutions) tightly in Parafilm to minimize evaporation, and store at 4
C until needed again for plating dilutions (for colony counting). 19. Incubate the square agar plate (lid-side-up) at 37
C until bacterial colonies are clearly visible (24–36 h). 20. When colonies are visible on the square agar plate, determine the appropriate dilutions for colony counting (i.e., the dilutions showing about 5–20 distinct colonies per spot), and label a 10 cm2 round LB agar plate for each treatment condition and dilution factor to be counted. 21. Retrieve the 96-well plate, transfer 100μL from each dilution to count onto the corresponding labelled round agar plate, and spread the culture evenly across the entire plate using an L-shaped sterile cell spreader. 22. Incubate the round agar plates (lid-side-down) at 37
C until bacterial colonies can be easily counted (24–36 h) (see Note 7). 23. Calculate CFU/mL by multiplying the number of CFU by the dilution factor, corrected for the amount of bacterial culture plated on the round agar plate. For example, if 100μL from the 104 dilution factor was spread onto a round agar plate, and 56 CFU were enumerated on that plate, the cell count is 56  105 CFU/mL. 24. Determine the MBC as the lowest antibiotic concentration for which the percentage of cells surviving antibiotic treatment (CFU/mL) dropped below 50%, relative to the CFU/mL of the mid-log phase culture (from Subheading 3.2, step 6). 3.3 Enrichment Method #1: Flow Sorting for Persisters by Staining Membrane Depolarization

Bacterial cells maintain an electrical potential gradient, also known as a “proton-motive force” (PMF), across their cytoplasmic membrane in order to synthesize the essential high-energy metabolite adenosine triphosphate (ATP). Multidrug tolerance in bacterial persistence has been associated with depolarized bacterial cells, which are seen to exhibit below-average levels of PMF [13– 15]. Fluorescence-activated cell sorting (FACS) can isolate persister cells from bacterial cultures treated with a PMF-sensitive dye. DiBAC4(3) is a bis-oxonal fluorescent dye that specifically permeates depolarized cells, binds to intracellular proteins or the cell membrane, and exhibits fluorescence with excitation and emission maxima at 490 and 516 nm, respectively.

Enrichment of Bacterial Persister Populations

77

While hyperpolarization of cells results in a rapid efflux of DiBAC4(3) and low fluorescence intensity (FI), cells with moderate-to-low PMF retain the dye and exhibit high FI. The final concentration of DiBAC4(3) used in the liquid culture media must be determined empirically to avoid dye-saturation. For Gramnegative bacteria, we found that a final concentration of 1μg/mL of DiBAC4(3) showed the best correlation between the percentage of antibiotic-surviving cells (persisters) and cells with high FI (low PMF). The high-DiBAC4(3) population can be sorted using the FACS technique described below. We found that a minimum of 106 sorted cells are required for lysis and subsequent bacterial RNA extraction. Collection of 106 sorted cells with high FI is a lengthy process (>4 h), given their low abundance (12 months can begin to lose their potency. 3. QIAzol contains two dangerous components: phenol and guanidine thiocyanate. QIAzol is toxic by inhalation, in contact with skin, and if swallowed. Handle inside a chemical fume hood. Contact with acids liberates highly toxic gas. Do not combine with bleach. Suitable fire extinguishing agents include CO2, extinguishing powder, water spray, and alcohol-resistant foam. Avoid inhaling gases liberated from explosion or combustion. 4. Colony stock plates that are used to make mid-log phase cultures must be discarded after 4 days to avoid genetic and/or phenotypic adaptation to low nutrient conditions. Plates should be wrapped tightly in Parafilm to minimize moisture loss. B.th. E264 colonies taken from plates >4 days old are sticky and difficult to resuspend into LB Broth. 5. Stationary phase liquid cultures (grown for 18 h) must be discarded after 8 h to avoid genetic and/or phenotypic adaptation to low nutrient conditions. Overnight cultures >8 h old can take longer to reach mid-log phase of growth when diluted in fresh LB Broth. 6. Freshly made mid-log phase cultures (OD600 ¼ 0.20–0.30) will continue to grow at ambient room temperature, even without constant agitation, for about 30 min after the 37
C incubation period. Cultures should therefore be removed from the shaking incubator about 30 min prior to treatment (as the OD600 approaches 0.20), then the OD600 should be measured again immediately prior to treatment. Do not allow mid-log phase cultures to overgrow, that is, OD600 > 0.30. 7. Do not allow colonies to grow very large, as neighboring colonies will begin to merge together and smaller colonies will be obscured; this makes it more difficult to obtain an accurate colony count. 8. Positive controls for DiBAC4(3) retention can also help to set the threshold gate for the collection of persisters with high FI levels. To make positive controls, use stationary-phase liquid cultures or bacteria treated with a metabolic inhibitor, then stain with DiBAC4(3) as described above. 9. Duration of antibiotic exposure must be empirically determined for each microbial species.

82

Samantha Adikari et al.

10. The low temperature allows for sufficient moisture retention within the agar medium to ensure cell viability throughout the aging period. 11. Freezing the cell lysates in QIAzol Lysis Reagent overnight improves the lysis efficiency and final RNA yield. For long-term storage, cell lysates should be stored at 80
C to prevent RNA degradation prior to the bacterial RNA extraction step.

Acknowledgments This work was supported by a Defence Threat Reduction Agency (DTRA) grant to EHG to study persistence and antibiotic tolerance in bacterial pathogens. References 1. Rocha-Granados MC, Zenick B, Englander HE et al (2020) The social network: Impact of host and microbial interactions on bacterial antibiotic tolerance and persistence. Cell Signal 75:109750 2. Van den Bergh B, Fauvart M, Michiels J (2017) Formation, physiology, ecology, evolution and clinical importance of bacterial persisters. FEMS Microbiol Rev 41:219–251 3. Gollan B, Grabe G, Michaux C et al (2019) Bacterial persisters and infection: past, present, and progressing. Annu Rev Microbiol 73:359–385 4. Leszczynska D, Matuszewska E, KuczynskaWisnik D et al (2013) The formation of persister cells in stationary-phase cultures of Escherichia coli is associated with the aggregation of endogenous proteins. PLoS One 8: e54737 5. Wood DN, Chaussee MA, Chaussee MS et al (2005) Persistence of Streptococcus pyogenes in stationary-phase cultures. J Bacteriol 187:3319–3328 6. Keren I, Minami S, Rubin E et al (2011) Characterization and transcriptome analysis of Mycobacterium tuberculosis persisters. mBio 2: e00100-00111 7. Keren I, Shah D, Spoering A et al (2004) Specialized persister cells and the mechanism of multidrug tolerance in Escherichia coli. J Bacteriol 186:8172–8180 8. Shah D, Zhang Z, Khodursky A et al (2006) Persisters: a distinct physiological state of E. coli. BMC Microbiol 6:53

9. Micheva-Viteva SN, Shakya M, Adikari SH et al (2020) A gene cluster that encodes histone deacetylase inhibitors contributes to bacterial persistence and antibiotic tolerance in Burkholderia thailandensis. mSystems 5:e00609 10. Micheva-Viteva SN, Ross BN, Gao J et al (2018) Increased mortality in mice following immunoprophylaxis therapy with high dosage of nicotinamide in Burkholderia persistent infections. Infect Immun 87:e00592–e00518 11. Ross BN, Micheva-Viteva S, Hong-Geller E et al (2019) Evaluating the role of Burkholderia pseudomallei K96243 toxins BPSS0390, BPSS0395, and BPSS1584 in persistent infection. Cell Microbiol 21:e13096 12. Nierman WC, Yu Y, Losada L (2015) The in vitro antibiotic tolerant persister population in Burkholderia pseudomallei is altered by environmental factors. Front Microbiol 6:1338 13. Meylan S, Porter CBM, Yang JH et al (2017) Carbon sources tune antibiotic susceptibility in Pseudomonas aeruginosa via tricarboxylic acid cycle control. Cell Chem Biol 24:195–206 14. Harms A, Maisonneuve E, Gerdes K (2016) Mechanisms of bacterial persistence during stress and antibiotic exposure. Science 354: aaf4268 15. Liu Z, Hu K, Tang Y et al (2018) SmpB downregulates proton-motive force for the persister tolerance to aminoglycosides in Aeromonas veronii. Biochem Biophys Res Commun 507:407–413

Part III Single-Cell Analysis of Persister Cells

Chapter 6 Observing Bacterial Persistence at Single-Cell Resolution Emma Dawson, Emrah S¸ims¸ek , and Minsu Kim Abstract Within a bacterial population, there can be a subpopulation of cells with an antibiotic-tolerant persister phenotype characterized by long lag phase. Their long lag phase necessitates long (hours or days) periods of single-cell observation to capture high-quality quantitative information about persistence. We describe a method of single-cell imaging using glass bottom dishes and a nutrient agarose pad that allows for longterm single-cell microscopy observation in a stable environment. We apply this method to characterize the lag phase and persistence of individual Escherichia coli cells. Keywords Persistence, Persisters, Single-cell microscopy, Lag time, Antibiotic resistance

1

Introduction When a bacterial population is exposed to antibiotics, a subpopulation called persisters may survive while the majority of cells quickly die. Single-cell imaging has revealed that persister cells exhibit long lag times when transferred from stationary phase to fresh nutrient medium. That is, while most of the population rejuvenates and enters a reproductive state, this subpopulation remains in the nondividing state for an extended period of time. We designed a set of experiments to allow one to observe persistence at single-cell resolution and characterize the effect of stationary phase on lag times. The first experiment, determining the time delay in killing of growing cells (Subheading 3.1), captures the distribution of the time delay in killing growing bacteria in the presence of an antibiotic. Cells are exposed to ampicillin and the time of killing is determined using single-cell microscopy. The second experiment, single cell observation of lag time after starvation (Subheading 3.2), captures the distribution of lag times of a previously starved bacterial population. Cells are starved of nutrients (nongrowing state) and then transferred to a nutrient-rich

Natalie Verstraeten and Jan Michiels (eds.), Bacterial Persistence: Methods and Protocols, Methods in Molecular Biology, vol. 2357, https://doi.org/10.1007/978-1-0716-1621-5_6, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2021

85

86

Emma Dawson et al.

environment. For each cell, the time at which it begins to grow (i.e., lag time) is determined using single-cell microscopy. Both experiments require the observation of single cells over long periods of time (hours or days) within a stable environment. To achieve this, we make use of glass bottom microscopy dishes, which allow cells to be sandwiched in the bottom of the dish between a nutrient agarose pad and a cover glass. The nutrient agarose pad is thick and enclosed to maintain a consistent environment around the cells for the duration of the experiment, dramatically increasing the possible duration of single-cell microscopy experiments. Experiments using this method follow a basic protocol: a nutrient agarose pad is prepared for the experiment, a liquid culture is sandwiched in a glass bottom dish beneath the nutrient agarose pad, the dish is mounted on a phase-contrast microscope, single-cell images are captured for ~20 h and the resulting images are analyzed using the MicrobeJ plugin for ImageJ. The specific protocols for each experiment are outlined in Subheadings 3.1 and 3.2, with in-depth descriptions of all the shared procedures provided in Subheadings 3.3–3.6.

2

Materials Prepare all solutions using nanopure water (prepared by purifying deionized water to attain a sensitivity of 18 MΩ·cm) and analytical grade reagents. Prepare and store all reagents at room temperature, unless otherwise indicated.

2.1 Determining Time Delay in Killing of Growing Cells

1. E. coli K12 NCM3772 ΔmotA strain, stored in 25% (v/v) glucose solution at
80  C. 2. Luria–Bertani (LB) broth: Add 600 mL water to a 1 L glass medium bottle with screw cap. Weigh 15 g of Luria–Bertani broth powder and transfer to the bottle. Close cap tightly and mix by inverting and swirling the bottle until the powder is completely dissolved. Loosen cap and autoclave at 121  C for 30 min. Allow the solution to cool before tightening cap. 3. 100 mg/mL ampicillin solution in water, stored at 4  C for up to a week. 4. 1 mM propidium iodide (PI) solution in water, prepare fresh. 5. Agarose powder. 6. 10 cm petri dish. 7. Glass culture tubes capable of holding 5 mL of culture and compatible caps (typical size is 20 mm). 8. 20 mL glass, screw-top scintillation vial. 9. Sterile P1000 pipette tips and compatible pipette.

Observing Bacterial Persistence at Single-Cell Resolution

87

10. Sterile 15 mL conical tube. 11. 35 mm glass bottom dish with a 20 mm well (see Note 1). 12. Metal scoopula (see Note 2). 13. Parafilm M laboratory film. 14. Optional: Cling film if storing pad overnight. 15. Spectrophotometer with standard cuvette. 16. Water bath, shaking at 250 rpm, 37  C. 17. Inverted microscope fitted with an automated mechanical XY stage, 60 objective, TRITC filter, and autofocus, housed in an incubator at 37  C. 18. ImageJ image analysis software with MicrobeJ plug-in (see Note 3). 2.2 Single-Cell Observation of Lag Time After Starvation

1. E. coli K12 NCM3772 ΔmotA strain, stored in 25% (v/v) glucose solution at
80  C. 2. N
C
minimal medium: Dissolve 1 g K2SO4, 13.5 g K2HPO4, 4.7 g KH2PO4, 0.048 g MgSO4, and 2.5 g NaCl in 1 L water and filter through a 0.22 μm pore MilliporeSigma filter. Store at 4  C (see Note 4). 3. Starvation medium: Mix 1 L N
C
minimal medium with 2.1 g ammonium chloride. Adjust pH to 7. 4. Glucose minimal medium: Mix 960 mL N
C
minimal medium with 2.1 g ammonium chloride and 40 mL of 1 M glucose solution in water. Adjust pH to 7. 5. Luria–Bertani (LB) broth: Add 600 mL water to a 1 L glass medium bottle with screw cap. Weigh 15 g of Luria–Bertani broth powder and transfer to the bottle. Close cap tightly and mix by inverting and swirling the bottle until the powder is completely dissolved. Loosen cap and autoclave at 121  C for 30 min. Allow the solution to cool before tightening cap. 6. 100 mg/mL ampicillin solution in water, stored at 4  C for up to a week. 7. Agarose powder. 8. 10 cm petri dish. 9. Glass culture tubes capable of holding 5 mL of culture and compatible caps (typical size is 20 mm). 10. 20 mL glass, screw-top scintillation vial. 11. Sterile P1000 pipette tips and compatible pipette. 12. Sterile 15 mL conical tube. 13. 35 mm glass bottom dish with a 20 mm well (see Note 1). 14. Metal scoopula (see Note 2).

88

Emma Dawson et al.

15. Parafilm M laboratory film. 16. Optional: Cling film if storing pad overnight. 17. Spectrophotometer with standard cuvette. 18. Water bath, shaking at 250 rpm, 37  C. 19. Inverted microscope fitted with an automated mechanical XY stage, 60 objective, and autofocus, housed in an incubator at 37  C. 20. ImageJ image analysis software, with MicrobeJ plug-in (see Note 3).

3

Methods

3.1 Determining Time Delay in Killing of Growing Cells

1. Prepare a seed culture by taking cells from a
80  C frozen stock and culturing in fresh LB overnight. 2. The following morning, transfer cells from the seed culture to fresh LB and culture for at least nine doublings, to an OD600 of 0.2–0.3 (experimental culture). 3. Following the protocol described in Subheadings 3.3 and 3.4, prepare an agarose pad of LB + 1% agarose + 100 μg/mL ampicillin + 4 μM PI (see Note 5) and use it to sandwich 4 μL of the experimental culture in a glass bottom dish. Seal the glass bottom dish with Parafilm and mount on a microscope with incubator set to 37  C. 4. Images of the cells should be captured in a time lapse using phase contrast and a TRITC filter in 10-min time intervals, and a duration of ~20 h. See Subheading 3.5. 5. Time-lapse images can then be analyzed in MicrobeJ, as described in Subheading 3.6, to determine the time of death for each cell. We found PI to be a good indicator of death by ampicillin (see Note 6), however, because it stains nucleic acids, it does not identify lysed cells. Thus, we measured the time at which each cell either was stained by PI or lost its phasecontrast refractivity as the time of death (since lysed cells lose their refractivity due to the loss of intracellular materials). Example images can be found in [1].

3.2 Single-Cell Observation of Lag Time After Starvation

1. Take cells from a
80  C frozen stock and culture in LB for 4–6 h (seed culture). 2. Transfer cells from the seed culture to glucose minimal medium at very low density (OD600 of ~0.0001 or below) and culture overnight (preculture) (see Note 7). 3. Dilute the preculture 20–50 into fresh glucose minimal medium and allow cells to grow exponentially for at least four doublings (experimental culture).

Observing Bacterial Persistence at Single-Cell Resolution

89

4. Wash the experimental culture. Begin by adding 5 mL of experimental culture to a prewarmed sterile 15 mL conical tube. 5. In a centrifuge prewarmed to 37  C, centrifuge the experimental culture at 1900 rcf for 5 min. 6. Discard supernatant and resuspend cells in 5 mL of starvation medium. 7. Repeat steps 5 and 6. 8. Transfer cells to a sterile glass culture tube and culture in starvation medium (starvation culture). 9. Following the protocol described in Subheadings 3.3 and 3.4, prepare an agarose pad with glucose minimal medium + 1% agarose + 100 μg/mL ampicillin and use it to sandwich 5 μL of the starvation culture in a glass bottom dish. Seal the glass bottom dish with Parafilm and mount the dish on a microscope with incubator set to 37  C. 10. Images of the cells should be captured in a time-lapse using phase-contrast exposure, 10-min time intervals, and a duration of ~20 h, see Subheading 3.5. 11. Time-lapse images can then be analyzed in MicrobeJ, as described in Subheading 3.6, to measure cell length in each frame. The lag time of each cell is then determined by the time after sandwiching at which its length begins to increase (see Note 8). 3.3 Preparing the Agarose Pad

1. Prewarm a rack of P100 pipette tips (see Note 9). 2. Add 3 mL of medium and 0.3 g agarose to a 20 mL glass, screw-top scintillation vial. Screw the cap loosely closed (to allow venting) and place the vial in a microwave. Warm the medium in short (2Chr gates for sorting the populations of interest. A representative gating strategy to minimize gate leakage is portrayed. Collect samples in flow cytometry tubes containing sterile spent medium, and divide each sorted population into treated and untreated subpopulations for your assay. 1Chr, 2Chr, and >2Chr indicate chromosome number (Chr). Abx antibiotic

2. Filter-sterilize the supernatant (spent medium) using an 0.22 μm filter into a 50 mL falcon tube. 3. Add 1 mL of sterile spent medium to labeled flow cytometry collection tubes, and bring these tubes to the sorter for cell

134

Allison M. Murawski et al.

collection. These tubes will be used to collect the sorted cells, such that the final volume in these tubes after sorting 1,000,000 cells will be 2 mL of 50% PBS/50% spent medium (see Note 14). 3.4 Controls for FACS: Determination of Whether the Timing Required to Sort Cells Using FACS Influences Persistence

Depending on the strain, the number of cells sorted, and the flow rate, the cell sorting process can take several hours or longer. For the conditions and sorter we used, sorting 1,000,000 wild-type cells took up to an hour in four-way purity sort mode. To determine whether this timing impacts the persistence phenotype of interest, we recommend conducting “presort” and “postsort” persister assay controls on unsorted cells. Briefly, dilute stained and unstained cells in 50% PBS/50% sterile spent medium to an OD600 that approximates the OD600 of the cells that were isolated from the sorter and collected in equal volume sterile spent medium. Subsequently, perform persister assays on these samples (see Subheading 3.7). “Presort” indicates that a persistence assay is initiated immediately prior to cell sorting, and “postsort” indicates that a persistence assay is initiated immediately after the conclusion of cell sorting. Both presort and postsort persistence assays should be performed on stained and unstained cells that are not sorted, and they are initiated at different points in time. The persister fractions of the presort condition can then be compared to those of the postsort condition to determine whether the timing required for cell sorting impacts persistence. These data can also be compared to the persister fraction arising from a Total Sort population to verify whether sorting itself influences persistence. See Subheading 3.7 for more details.

3.5

FACS enables the segregation of bacteria into smaller subpopulations based on a phenotypic characteristic that is quantitatively measured via a fluorescence value. Follow your manufacturer’s guidelines for proper sorter start-up, alignment/QC, and sort setup. For each FACS experiment, take special consideration of the biosafety level of the organism you work with; FACS functions under a pressured system which can produce aerosols during a sort failure, such as a nozzle clog. Accidental inhalation of these droplets can cause unintentional risks to your health. A proper risk assessment should be performed with your institutional biosafety officer prior to conducting sort experiments and all safety recommendations should be implemented. Additionally, proper precautions should be adhered to in order to ensure that the FACS sorter remains sterile and free from particulate matter (see Note 15). Prior to and after each sort experiment, the sorter should be cleaned with a freshly prepared 10% bleach solution for 10 min, followed by a 10 min wash with sterile PBS or autoclaved DI water. Weekly sterilizations of the sheath fluid container via autoclave and weekly sterilization of the fluid lines with 70% ethanol are

FACS

Counting Chromosomes in Persisters

135

recommended, as per your manufacturer’s instructions. Frequent incubation of sheath fluid captured into a tube from above the waste catcher on an LB agar plate at 37
C for 16–24 h is recommended. A lack of colony growth confirms a sterile sorter. Regular maintenance on the sorter, including laser alignment, with a particular focus on the UV laser if Hoechst 33342 is to be utilized, is recommended. 1. Turn on the FACS sorter ~30 min early to allow the sorter and UV laser to warm up, according to the manufacturer’s instructions (see Note 16). 2. Set the laser delay values, area scaling factors, droplet breakoff position, and drop delay values according to the instrument’s instructions. 3. Obtain single-cell Hoechst 33342 fluorescent values by exciting each cell with a 355 nm laser run at 60 mW power (Coherent, Santa Clara, CA) and filtering emitted light through a 410 nm long pass filter and a 450/50 nm bandpass filter. 4. If your sorter has temperature control options, select a temperature at which you wish to sort the bacteria (see Note 17). 5. Select the appropriate sorter nozzle and pressure for sorting bacteria (see Note 18). 6. To place area measurements on the same scale as height measurements for each sorting session, optimize forward scatter (FSC) and UV area scaling factors using 1 μm Yellow-Green fluorescent beads (see Note 19). 7. Set FSC and side scatter (SSC) parameters to be recorded on a log scale. Adjust Hoechst 33342 fluorescence values to be recorded on a linear scale (see Note 20). To identify cells: 8. Run a sample of clean, filtered PBS to assess particulate contamination in tubing and electronic noise. Adjust FSC and SSC voltages and threshold values to minimize electronic noise. 9. Identify bacterial cells using FSC and SSC as determined by an unstained control (see Note 21) and adjust voltages to visualize the entire population. 10. Exclude cell aggregates by gating first on SSC-H and SSC-W, followed by gating with Hoechst 33342 as determined by Hoechst 33342-W and Hoechst 33342-A (see Note 22). 11. Place a sample of unstained, live stationary-phase E. coli on the sorter and adjust voltages such that electronic noise is minimized and cells are on scale (see Note 23). Record a data file of at least 50,000 cells.

136

Allison M. Murawski et al.

12. Backflush the system sample line to ensure unstained cells are removed from the tubing. 13. Place the stained sample on the sorter and adjust both doublet discrimination gates if necessary. Set gates representative of the one-chromosome population and the two-chromosome population, as demonstrated in Fig. 2. Employ a conservative gating strategy to ensure as minimal leakage between the gates as possible. 14. Sort 1,000,000 cells per gate (see Subheading 3.7), and collect cells in 1 mL of sterile spent medium in a flow cytometry tube (these tubes were prepared in Subheading 3.3.2). After sorting, place the cells back in the dark container. 15. For a Total Sort sample, sort 1,000,000 cells as determined by the SSC-H and SSC-W and the Hoechst 33342-W and Hoechst 33342-A gates. Collect cells in 1 mL of sterile spent medium in a flow cytometry tube. 3.6 Verification of Sorting Fidelity Based on DNA Content

3.6.1 Confirmation of DNA Content of Sorted Cells Using a Secondary dsDNA-Specific Dye (PicoGreen)

To assess the fidelity of the sorting methodology, we recommend using two methods to evaluate sorting precision. First, we suggest restaining sorted cells with a secondary dsDNA-specific dye and analyzing the restained subpopulations using flow cytometry. In our experiments, we restained the sorted Hoechst-stained subpopulations with PicoGreen, a known cell-impermeable dsDNA-specific dye (see Note 24). Second, to quantify sort purity using a method independent of staining, we recommend sorting a strain where chromosome copy number is reported by a fluorescent protein. In our experiments, we used an origin reporter strain that has distinct fluorescent foci indicative of the number of chromosomes within single cells, and visualized the subpopulations using confocal fluorescence microscopy. Both methods are described herein. 1. Grow bacteria to stationary phase as described in Subheading 3.1. 2. After 20 h, remove 1 mL, spin for 3 min at 15,000 rpm in a microcentrifuge tube, and fix in 1 mL of 4% PFA for 30 min at room temperature (see Note 25). 3. Centrifuge the cells for 3 min at 15,000 rpm. 4. Resuspend in 1 mL of PBS. 5. Stain the cells with 5 μg/mL of Hoechst 33342 and prepare for FACS, as described in Subheading 3.3. 6. Sort 1,000,000 cells per gate based on Hoechst 33342 fluorescence as described in Subheading 3.5. Collect in empty flow cytometry tubes.

Counting Chromosomes in Persisters

137

7. Permeabilize the sorted cells with 9 mL of 70% ethanol in a 15 mL polypropylene conical tube for at least 3 h at 4
C (see Note 26). 8. Spin the permeabilized cells at 4000 rpm for 10 min at 4
C. 9. Remove 9 mL of supernatant. 10. Resuspend the pellet and transfer the remaining ~1 mL of cells to a microcentrifuge tube. 11. Spin for 3 min at 15,000 rpm. 12. Remove the supernatant. 13. Leave the tubes open overnight and cover with low lint tissue to allow residual ethanol to evaporate. 14. The next morning, resuspend the dried pellet in 500 μL of PBS. 15. Dilute a PicoGreen dimethyl sulfoxide (DMSO) stock 1:100 in 25%/75% DMSO/MilliQ water. Stain the ethanol-fixed cells in PBS with 100 μL of the diluted PicoGreen solution or 100 μL of solvent for an unstained control. Incubate in the dark at room temperature for ~3 h. 16. Centrifuge the cells for 3 min at 15,000 rpm. 17. Remove 500 μL of supernatant. 18. Resuspend the cell pellet in the remaining 100 μL and transfer to a de-capped PCR tube placed into a flow cytometry tube. 19. Acquire single-cell flow cytometric data: (a) Turn on the cytometer ~30 min early to allow the cytometer and lasers to warm up and perform a QC check as per manufacturer’s instructions (see Note 27). (b) Identify single cells using FSC and SSC and perform doublet discrimination as described in Subheading 3.5. Set the PicoGreen channel to linear scaling. (c) Apply no gates to the sort-check samples (see Note 28). (d) Acquire single-cell PicoGreen fluorescence values by exciting each cell with a 488 nm laser run at 100 mW power (Coherent, Santa Clara, CA) and collect using a 505 nm long pass filter and a 525/50 nm bandpass filter. Evaluate fidelity of sorted populations. 3.6.2 Confirmation of DNA Content of Sorted Cells Using an Origin-of-Replication Reporter

1. Grow origin reporter strain ori1-parS + pALA2705 in M9 minimal glucose medium in the presence of 100 μg/mL of AMP for plasmid retention as described in Subheading 3.1 (see Note 29). 2. Fix the cells with 4% PFA as described in steps 2–4 of Subheading 3.6.1.

138

Allison M. Murawski et al.

3. Stain with 5 μg/mL of Hoechst 33342 as described in Subheading 3.3.1. 4. Using FACS (Subheading 3.5), sort 2,000,000 of one-chromosome, two-chromosome, and Total Sort subpopulations each into empty flow cytometry tubes (see Note 30). 5. Centrifuge the eluted cells for 3 min at 15,000 rpm and remove all but ~30 μL of supernatant. Store in the dark at 4
C until microscopic analysis. 6. On the day of microscopic analysis, dissolve 0.1 g of agarose in 10 mL of autoclaved MilliQ water by microwaving for ~1 min in a 50 mL falcon tube. 7. After cooling slightly, pipet 1 mL of agarose solution onto sterile plain microscopy slides. 8. Lay a second plain microscopy slide cross-wise over top by supporting with adjacent double-stacked plain microscopy slides laid parallel to and on both sides of the original slide. 9. Allow to harden at room temperature for ~1 h. 10. Remove the top slide carefully, and cut the agarose pad into a ~2  2 cm2. 11. Pipet 2 μL of the fixed, sorted, and concentrated ori1parS + pALA2705 cell suspensions onto the agarose pad and allow to air dry. 12. Pipet another 2 μL of cell suspension on top of the original spot to increase cell density, allow to air dry. Repeat a third time. 13. Overlay a cover glass and seal with clear nail polish. 3.6.3 Controls for Origin Reporter

1. Grow MG1655 with pALA2705 (no parS binding site) and ori1-parS with pALA2705ΔparB (contains parS binding site and expresses GFP but lacks ParB) to stationary phase as described in Subheading 3.1. Use 100 μg/mL of AMP for plasmid retention. 2. Fix in 4% PFA as described in steps 2–4 of Subheading 3.6.1 and resuspend in PBS for microscopic analysis. 3. Dilute these samples 1:12 in PBS. 4. Prepare agarose microscopy slides as described in steps 6–10 in Subheading 3.6.2. 5. Spot 2 μL onto the agarose and allow to air dry. 6. Overlay with a cover glass and seal with clear nail polish.

3.6.4 Confocal Microscopy

1. Visualize the origin reporter strain using confocal fluorescence microscopy (see Note 31). We recommend observing the bacteria using phase contrast and visualizing the foci by taking a Z-stack using the 488 nm laser (see Notes 32–34). A

Counting Chromosomes in Persisters

139

Fig. 3 Representative phase contrast image with fluorescent overlay of origin reporter strain ori1-parS with the parS site located at position 3928826. Scale bar represents 5 μm

representative confocal fluorescence microscopy image of an unsorted, stationary-phase origin-reporter strain is portrayed in Fig. 3. 2. Use image analysis software, such as MicrobeJ, to count foci within single cells (see Note 35), and correlate this data with the PicoGreen sort check data (Subheading 3.6.1). 3.7 Culturability and Persistence Assays on Cell Samples

In this section, you will want to use persister fractions measured from the assays conducted in Subheading 3.2.2 to estimate the number of persisters you would expect to observe from the number of cells you choose to sort. You can then calculate the estimated number of persisters you expect to count at each dilution when plating serial dilutions, keeping in mind the number and volume of cells in the assay, whether the sample is divided into treated and untreated conditions, and the volume of cells plated for CFU measurements. If taking hourly time points where survival is higher during earlier time points, you may also want to plate additional dilutions. If your persister fraction is very low, we recommend only taking initial and final CFUs (no intermediate time points), and plating all remaining volume at the final time point. Controls with cells that were not sorted but treated under identical conditions as sorted cells can assess whether biphasic survival curves are attained under those treatment conditions.

3.7.1 Culturability and Persistence Assay on Sorted Fractions

Once the cells have been sorted based on chromosomal content, proceed with assays that address your question of interest. As an example, we provide instructions on how to perform persistence assays on the sorted subpopulations.

140

Allison M. Murawski et al.

1. Divide each sorted sample from Subheading 3.5 (which, in our case, is 2 mL in total volume) into two 1 mL aliquots in 15 mL polystyrene tubes to represent treatment and control conditions. 2. Remove 100 μL of cells from each tube and add to microcentrifuge tubes (initial samples for initial CFU/mL quantifications). 3. Treat each culture (step 1) with your antibiotic of choice or an equal volume control solvent. 4. Incubate cultures at 37
C with shaking at 250 rpm for 5 h. 5. Process initial samples by adding 900 μL of PBS to the microcentrifuge tubes, centrifuging at 15,000 rpm for 3 min, and removing 900 μL of supernatant. Resuspend the pellet in the remaining 100 μL of PBS, and perform serial dilutions as described in step 10 of Subheading 3.1. Spot 10 μL of each dilution for the first six wells onto LB agar, and incubate the agar plates for 16–24 h at 37
C. 6. At various time points during the assay, remove 100 μL of cell culture from each tube and add to microcentrifuge tubes. 7. Add 900 μL of PBS. 8. Spin at 15,000 rpm for 3 min. 9. Remove 900 μL of supernatant. 10. Repeat steps 7–9 until the residual antibiotic concentration in the treated cultures is below the MIC (see Note 11) and perform the same number of washes on the untreated controls. 11. Resuspend the pellet in the remaining 100 μL. For the untreated samples, perform serial dilutions and plate 10 μL of each dilution onto LB agar and allow to dry. For the treated samples, process identically and plate 50 μL of each dilution, instead of 10 μL to increase the limit of detection (see Note 36). 12. Incubate the agar plates for 16–24 h at 37
C, and count CFUs at the appropriate dilution (between 20–100 CFUs) the following day. 3.7.2 Culturability and Persistence Assays on Unsorted Cells Diluted to the Same Density of Sorted Cells

These steps can be conducted in parallel with the sorted cell persister assays described in Subheading 3.7.1. 1. To demonstrate that FACS sorting does not alter the persistence phenotype, dilute the stained and unstained cells samples that had been FACS analyzed and/or sorted to an OD600 < 0.001 in 2 mL of 50%/50% PBS/sterile spent medium. 2. Follow steps 1–12 in Subheading 3.7.1.

Counting Chromosomes in Persisters

4

141

Notes 1. Other fluorescent dyes and fluorescent proteins that form foci can be used to quantify chromosomal content. However, one should ensure that the dyes are nontoxic and that foci are produced within the vast majority of cells. 2. Any medium can be used, but we note that the rate of growth within the culture may affect chromosome abundances. 3. We note that any antibiotic can be used for persistence assays, and recommend that the antibiotic concentration be high enough such that persister quantification occurs in the concentration-independent portion of survival plots (regime where modest changes to antibiotic concentration do not alter survival appreciably). For the persistence assay portrayed in Fig. 1, a 5 mg/mL stock of MOXI was made by dissolving it in autoclaved MilliQ water until fully dissolved, and samples were treated at a final concentration of 5 μg/mL. Further, biphasic killing kinetics must be observed for treatments to yield persister measurements. 4. We recommend using medium that was freshly prepared less than a week beforehand. In addition, the length of time to reach stationary phase varies with different strains and different bacteria; 20 h is a suggestion and works well for E. coli K12 MG1655, but can be shortened or extended as necessary. 5. We generally dilute 10 μL of cells into 290 μL of PBS in a flatbottom 96-well plate before measuring the OD600 using a spectrophotometer. We calculate the OD600 by accounting for the 30-fold dilution. 6. For CFU enumeration, we recommend drying LB agar plates covered and at room temperature for ~2 days prior to use in order to prevent 10 μL spots from running together. 7. We generally count dilutions that contain 20–100 bacteria. 8. Fluorescent DNA binding proteins, such as HU-GFP, have also been previously used to assess DNA content within single cells [43, 44]. 9. We observed that the age and method of storage for antibiotic solutions can affect killing kinetics, especially for persistence assays. We therefore recommend making antibiotic stocks fresh on the day of the experiment. 10. Additional time points beyond 5 h of treatment may be required to see biphasic killing kinetics. 11. Calculate the MIC for each strain tested using agar or broth dilution methods [45], or an antibiotic-specific Etest strip (bioMe´rieux, Marcy-l’E´toile, France).

142

Allison M. Murawski et al.

12. Check the sorter manufacturer’s guidelines for appropriate cell density for cell sorting. We determined that 2.5  107 to 3  107 cells/mL is a good density for cell sorting with the machine used here (see Note 16). 13. We used prewarmed heating packs placed in Styrofoam containers to transport cells at desired temperatures. 14. Provide extra empty flow cytometry tubes and flow cytometry tubes containing filter-sterilized spent medium as backup. 15. To assess the sorter’s sterility prior to sorting, when arriving at the sorter obtain a sample of sheath fluid directly from the sheath fluid stream by catching several mL into a clean tube and plating directly onto LB agar. Sterility should be observed after incubation at 37
C for 16–24 h. 16. We performed sorting with a BD Biosciences FACSAria Fusion Special Order Research Product (SORP) (San Jose, CA) system with a 70 μm nozzle at 70 psi and window extension 10.0. We used FACSDiva version 8.0 software to acquire and analyze the data. Since each FACS sorter is unique, we strongly recommend testing and optimizing specific settings prior to conducting all sort experiments. 17. Per our sorter’s capabilities, we kept our sort sample and collection tubes at 37
C. 18. We used a 70 μm nozzle and sorted bacteria at 70 psi system pressure using the FACSAria Fusion. 19. Adjust the area scaling factors for the height (H) and area (A) signals for each detector until the median signal intensity values of the corresponding H and A signals are equivalent in order to place area measurements on the same scale as height measurements. 20. Forward scatter (FSC) is relatively proportional to a particle’s cross-sectional area and refractive index, whereas side scatter (SSC) is relatively proportional to a particle’s internal complexity or granularity. Small particles such as bacteria are best visualized with FSC and SSC parameters set to a log scale. We used a linear scale to observe DNA content via Hoechst 33342 staining to evenly space fluorescence peaks along the x-axis of a histogram, which correlate with integer numbers of chromosomes. 21. With our conditions, we set signal thresholds to SSC 200 and DAPI 1000 channels for Hoechst 33342–stained cells and to SSC 200 and DAPI 200 channels for unstained cells. The threshold values need to be optimized on all instruments to ignore electronic noise and record data for all bacterial cells. 22. For DNA content analysis where we want to observe a twofold change in DNA content, it is important to create doublet

Counting Chromosomes in Persisters

143

discrimination plots and reduce the likelihood of sorting cells that are clumped together. For example, we want to ensure that a diploid signal we observe is due to the presence of a cell harboring two chromosomes, and not because two monoploid cells are stuck together. To do so, we recommend gating on SSC-H and SSC-W, followed by gating with Hoechst 33342 as determined by Hoechst 33342-W and Hoechst 33342-A (Fig. 2a). 23. Monitoring the Threshold Rate (events per second) gives an indication of the amount of electronic noise present based on the parameter threshold and voltages selected. As FSC and SSC voltages are adjusted, the events/second rate may greatly increase, indicating that electronic noise is surpassing the threshold value(s). 24. Both dsDNA-specific dyes should have distinct fluorescent properties (e.g., unique excitation and emission spectra). For concomitant use, the dye and the fluorescent protein should also be fluorometrically distinct. 25. For verification of sorting fidelity, fixation is not required and live cell populations can instead be used. We fixed the cells for convenience, so as to not have to rush to permeabilize the cells for staining with PicoGreen. 26. While the cells are already fixed in PFA before Hoechst 33342 staining, PicoGreen requires permeabilization with ethanol to access and bind to the intracellular DNA. Both fixation methods are required when using these fluorescent dyes together. 27. We used an LSRII SORP (Special Order Research Product) flow cytometer and Cytometer Setup and Tracking Beads for QC (BD Biosciences, San Jose, CA). 28. Since we are conducting a purity check, no gates should be applied to the reanalyzed samples. However, doublet discrimination gates may be necessary if cells have clumped during preparation steps between sorting and flow cytometry. 29. For strain ori1-parS + pALA2705, induction with IPTG is not required to observe foci, as previously determined [22]. 30. We recommend increasing the number of cells sorted when subsequently visualizing these cells using confocal microscopy in order to increase cell density and ensure multiple cells can be visualized in each frame. 31. We used a Nikon Ti-E (Nikon Instruments, Melville, NY) equipped with a Yokogawa spinning disc (CSU-21) (Yokogawa, Tokyo, Japan) mounted with a quad dichroic beam splitter accommodating 405, 488, 561, and 647 lasers (Chroma, Bellows Falls, VT). 32. In our setup, the microscope is controlled using NIS-Elements, V4.5 (Nikon Instruments). The spinning disc

144

Allison M. Murawski et al.

detector is a Hamamatsu ImagEM back-thinned EMCCD (Hamamatsu, Tokyo, Japan). We use an emission filter, ET525/50m (Chroma) for emission discretion after the spinning disc. We use an Agilent laser launch with a 488 nm laser for fluorescence imaging and a piezo stage with 100 μm range capable up to 100 steps/s. We use a CF160 Plan Fluor Phase Contrast DLL 100 oil 1.3NA objective with corresponding annulus. 33. For our purposes, exposure for phase contrast images is set to 300 ms, and 1 s exposure at 50% laser power is used for fluorescence imaging. Exposure and laser power are optimized to limit photobleaching and maximize the signal to noise ratio. 34. Since the origin-of-replication foci within the cells may not be within the same plane, we recommend taking GFP Z-stacks every 0.3 μm for 21 steps. 35. We note that while the origin reporter is a useful tool to quantify chromosomes in untreated and fixed bacteria, certain conditions may cause destruction of the foci, leading to diffuse fluorescence within the cells. If intending to quantify chromosomes, we recommend testing any experimental conditions on the origin reporter strain first to see if the treatment alters foci formation and chromosome abundance enumeration. 36. The plating volume can be adjusted in order to observe 20–100 CFUs at your expected dilution.

Acknowledgments This work was supported by the National Institute of Allergy and Infectious Disease of the National Institutes of Health (A.M.M: F30AI140697; M.P.B.: R01AI130293). The Princeton University Flow Cytometry Resource Facility is supported, in part, with funding from NCI-CCSG P30CA072720-5921. This work is the responsibility of the authors and the funders have no role in the design or implementation of the experiments or the decision to publish. References 1. Balaban NQ, Helaine S, Lewis K et al (2019) Definitions and guidelines for research on antibiotic persistence. Nat Rev Microbiol 17:441–448 2. Balaban NQ, Merrin J, Chait R et al (2004) Bacterial persistence as a phenotypic switch. Science 305:1622–1625 3. Keren I, Kaldalu N, Spoering A et al (2004) Persister cells and tolerance to antimicrobials. FEMS Microbiol Lett 230:13–18

4. Barrett TC, Mok WWK, Murawski AM, Brynildsen MP (2019) Enhanced antibiotic resistance development from fluoroquinolone persisters after a single exposure to antibiotic. Nat Commun 10:1177 5. Fisher RA, Gollan B, Helaine S (2017) Persistent bacterial infections and persister cells. Nat Rev Microbiol 15:453–464 6. Wilmaerts D, Windels EM, Verstraeten N, Michiels J (2019) General mechanisms leading

Counting Chromosomes in Persisters to persister formation and awakening. Trends Genet 35:401–411 7. Helaine S, Kugelberg E (2014) Bacterial persisters: formation, eradication, and experimental systems. Trends Microbiol 22:417–424 8. Amato SM, Fazen CH, Henry TC et al (2014) The role of metabolism in bacterial persistence. Front Microbiol 5:70 9. Mok WWK, Brynildsen MP (2018) Timing of DNA damage responses impacts persistence to fluoroquinolones. Proc Natl Acad Sci U S A 115:E6301–E6309 10. Akerlund T, Nordstro¨m K, Bernander R (1995) Analysis of cell size and DNA content in exponentially growing and stationary-phase batch cultures of Escherichia coli. J Bacteriol 177:6791–6797 11. Hill NS, Kadoya R, Chattoraj DK, Levin PA (2012) Cell size and the initiation of DNA replication in bacteria. PLoS Genet 8: e1002549 12. Kraemer J, Sanderlin A, Laub M (2019) The stringent response inhibits DNA replication initiation in E. coli by modulating supercoiling of oriC. mBio 10:e01330 13. Soppa J (2017) Polyploidy and community structure. Nat Microbiol 2:1–2 14. Soppa J (2015) Polyploidy in archaea and bacteria: about desiccation resistance, giant cell size, long-term survival, enforcement by a eukaryotic host and additional aspects. J Mol Microbiol Biotechnol 172730(24):409–419 15. Michelsen O, Hansen FG, Albrechtsen B, Jensen PR (2010) The MG1363 and IL1403 laboratory strains of Lactococcus lactis and several dairy strains are diploid. J Bacteriol 192:1058–1065 16. Krisko A, Radman M (2013) Biology of extreme radiation resistance: the way of Deinococcus radiodurans. Cold Spring Harb Perspect Biol 5:a012765 17. Ohtani N, Tomita M, Itaya M (2010) An extreme thermophile, Thermus thermophilus, is a polyploid bacterium. J Bacteriol 192:5499–5505 18. Krasin F, Hutchinson F (1977) Repair of DNA double-strand breaks in Escherichia coli, which requires recA function and the presence of a duplicate genome. J Mol Biol 116:81–98 19. Soler-Bistue´ A, Mondotte JA, Bland MJ et al (2015) Genomic location of the major ribosomal protein gene locus determines Vibrio cholerae global growth and infectivity. PLoS Genet 11:e1005156 20. Couturier E, Rocha E (2006) Replicationassociated gene dosage effects shape the genomes of fast-growing bacteria but only for

145

transcription and translation genes. Mol Microbiol 59:1506–1518 21. Sun L, Alexander HK, Bogos B et al (2018) Effective polyploidy causes phenotypic delay and influences bacterial evolvability. PLoS Biol 16:e2004644 22. Espeli O, Mercier R, Boccard F (2008) DNA dynamics vary according to macrodomain topography in the E. coli chromosome. Mol Microbiol 68:1418–1427 23. Li Y, Austin S (2002) The P1 plasmid is segregated to daughter cells by a ‘capture and ejection’ mechanism coordinated with Escherichia coli cell division. Mol Microbiol 46:63–74 24. Aedo SJ, Orman MA, Brynildsen MP (2019) Stationary phase persister formation in Escherichia coli can be suppressed by piperacillin and PBP3 inhibition. BMC Microbiol 19:140 25. Ferullo DJ, Lovett ST (2008) The stringent response and cell cycle arrest in Escherichia coli. PLoS Genet 4:e1000300 26. Ferullo DJ, Cooper DL, Moore HR, Lovett ST (2009) Cell cycle synchronization of E. coli using the stringent response, with fluorescence labeling assays for DNA content and replication. Methods 48:8–13 27. Marie D, Vaulot D, Partensky F (1996) Application of the novel nucleic acid dyes YOYO-1, YO-PRO-1, and PicoGreen for flow cytometric analysis of marine prokaryotes. Appl Environ Microbiol 62:1649–1655 28. Hoechst 33342, trihydrochloride, trihydrate— 10 mg/mL solution in water. https://www. thermofisher.com/order/catalog/product/ H3570. Accessed 20 Nov 2020 29. Nucleic acid stains—section 8.1—US. https:// www.thermofisher.com/us/en/home/ references/molecular-probes-the-handbook/ nucleic-acid-detection-and-genomics-technol ogy/nucleic-acid-stains.html. Accessed 20 Nov 2020 30. Little D, Luft C, Mosaku O et al (2019) Highcontent analysis of mitochondrial function in iPSC-derived neurons. In: Mandenius C-F, Ross JA (eds) Cell-based assays using iPSCs for drug development and testing. Springer, New York, NY 31. Lee D, Ko W-K, Hwang D-S et al (2016) Use of baicalin-conjugated gold nanoparticles for apoptotic induction of breast cancer cells. Nanoscale Res Lett 11:381 32. Wang R, Dang Y-L, Zheng R et al (2014) Live cell imaging of in vitro human trophoblast syncytialization. Biol Reprod 90:117 33. Belloc F, Dumain P, Boisseau MR et al (1994) A flow cytometric method using Hoechst 33342 and propidium iodide for simultaneous

146

Allison M. Murawski et al.

cell cycle analysis and apoptosis determination in unfixed cells. Cytometry 17:59–65 34. Rice GC, Gray JW, Dean PN, Dewey WC (1984) Fluorescence-activated cell sorting analysis of the induction and expression of acute thermal tolerance within the cell cycle. Cancer Res 44:2368–2376 35. Srivastava VK, Nalbantoglu J (2008) Flow cytometric characterization of the DAOY medulloblastoma cell line for the cancer stemlike phenotype. Cytometry A 73A:940–948 36. Blaheta RA, Kronenberger B, Woitaschek D et al (1998) Development of an ultrasensitive in vitro assay to monitor growth of primary cell cultures with reduced mitotic activity. J Immunol Methods 211:159–169 37. van Zandvoort MAMJ, de Grauw CJ, Gerritsen HC et al (2002) Discrimination of DNA and RNA in cells by a vital fluorescent probe: lifetime imaging of SYTO13 in healthy and apoptotic cells. Cytometry 47:226–235 38. Knowles RB, Sabry JH, Martone ME et al (1996) Translocation of RNA granules in living neurons. J Neurosci 16:7812–7820 39. Poot M, Gibson LL, Singer VL (1997) Detection of apoptosis in live cells by MitoTracker™ Red CMXRos and SYTO dye flow cytometry. Cytometry 27:358–364

40. Wlodkowic D, Skommer J, Pelkonen J (2007) Towards an understanding of apoptosis detection by SYTO dyes. Cytometry A 71A:61–72 41. Guindulain T, Comas J, Vives-Rego J (1997) Use of nucleic acid dyes SYTO-13, TOTO-1, and YOYO-1 in the study of Escherichia coli and marine prokaryotic populations by flow cytometry. Appl Environ Microbiol 63:4608–4611 42. Hirons GT, Fawcett JJ, Crissman HA (1994) TOTO and YOYO: new very bright fluorochromes for DNA content analyses by flow cytometry. Cytometry 15:129–140 43. Goormaghtigh F, Melderen LV (2019) Singlecell imaging and characterization of Escherichia coli persister cells to ofloxacin in exponential cultures. Sci Adv 5:eaav9462 44. Wery M, Woldringh CL, Rouviere-Yaniv J (2001) HU-GFP and DAPI co-localize on the Escherichia coli nucleoid. Biochimie 83:193–200 45. Wiegand I, Hilpert K, Hancock REW (2008) Agar and broth dilution methods to determine the minimal inhibitory concentration (MIC) of antimicrobial substances. Nat Protoc 3:163–175

Part IV Omics Approaches in Persistence Research

Chapter 10 Analyzing Persister Proteomes with SILAC and Label-Free Methods Bork A. Berghoff Abstract State-of-the-art mass spectrometry enables in-depth analysis of proteomes in virtually all organisms. This chapter describes methods for the analysis of persister proteomes by mass spectrometry. Stable isotope labeling by amino acids in cell culture (SILAC) is applied to assess protein biosynthesis in persister cells, which are isolated by treatment with beta-lactam antibiotics. Furthermore, persister proteomes during the postantibiotic recovery phase are analyzed by label-free quantification. The presented methods are valuable tools to shed light on persister physiology. Key words Proteomics, Mass spectrometry, SILAC, Label-free quantification, Persister cells, Postantibiotic recovery

1

Introduction Persister cells are in a transient state of reduced cellular activity, which allows them to withstand antibiotic treatments for an extended period compared to the actively growing part of the population. Assessing persister physiology is key to understanding which factors determine the persister state. Persister physiology may be studied at several levels, such as morphology, metabolic state and gene expression. High-throughput methods provide global information on molecules with changed abundances, and have been proven useful to study persister subpopulations. An important step for high-throughput analyses is isolation of persister fractions, which can be achieved by fluorescence-activated cell sorting (FACS) of reporter strains [1, 2]. Alternatively, persister fractions can be isolated by elimination of nonpersisters using betalactam antibiotics during exponential phase [3], which is considered as a simple and convenient method. However, wild-type cultures suffer from low persister levels in exponential phase, which complicates persister isolation and high-throughput analyses.

Natalie Verstraeten and Jan Michiels (eds.), Bacterial Persistence: Methods and Protocols, Methods in Molecular Biology, vol. 2357, https://doi.org/10.1007/978-1-0716-1621-5_10, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2021

149

150

Bork A. Berghoff

Controllable expression of toxins or administration of bacteriostatic agents can be applied to induce the persister state and to increase the size of persister fractions [4, 5]. Alternatively, mutants with a “high persistence” (hip) phenotype are ideal model systems to apply high-throughput methods [3]. High-throughput proteomics by mass spectrometry (MS) analysis is a powerful tool, but has rarely been addressed to study persister physiology in planktonic cultures [6]. However, during the last years this gap was closed by several studies, using different persister enrichment and isolation methods [7–10]. The methods described here were used to assess the persister proteome of a high persistence mutant in Escherichia coli [11]. The type I toxin–antitoxin system TisB/IstR-1 is involved in persister formation upon DNA damage stress [12]. TisB is a small protein that targets the inner membrane, causing depolarization and ATP depletion [13, 14]. Deregulation of toxin TisB by deletion of regulatory RNA elements produced mutant Δ1–41 ΔistR, which has an enriched persister fraction during exponential phase [15]. Persister cells were isolated by treatment with the beta-lactam antibiotic ampicillin during exponential phase. In the Methods section, two different protocols for analysis of persister proteomes will be described: (1) analysis of protein biosynthesis in persister cells by metabolic labeling using stable isotope amino acids (SILAC), and (2) analysis of the persister proteome during postantibiotic recovery by label-free quantification. SILAC was originally developed for mammalian cell lines [16], but has successfully been applied to several microorganisms [17– 20]. Importantly, auxotrophic strains are not mandatory for bacterial SILAC experiments, since incorporation of stable isotope-containing amino acids is sufficiently high (92%) also in prototrophic strains [17, 18]. In a pulsed-SILAC approach, the stable isotope amino acid L-lysine-13C6,15N2 (Lys-8) was administered in parallel to the ampicillin treatment during exponential phase (Fig. 1a). Using liquid chromatography (LC)-MS/MS analysis, protein biosynthesis in persister cells was assessed by Lys-8 incorporation into proteins. In exponential cultures without ampicillin, Lys-8 incorporation was on average 71.9% after 4 h of incubation (Fig. 1b). In persister fractions, isolated by ampicillin, Lys-8 incorporation was on average 33.6%, demonstrating that persister cells have a reduced translational activity and/or protein turnover (Fig. 1c). Using Significance B analysis [21], several proteins were identified in the persister fraction with a Lys-8 incorporation significantly higher than the average of proteins (Fig. 1d). These proteins are likely physiological markers of persister cells, as exemplified by the master regulator of the general stress response, RpoS, and several RpoS-dependent proteins [11]. Another intriguing question in the persister field concerns persister awakening and postantibiotic recovery. Isolated persister fractions from the high persistence mutant Δ1–41 ΔistR were

Fig. 1 Pulsed-SILAC and LC-MS/MS to assess protein biosynthesis in persister cells. (a) Experimental workflow of the pulsed-SILAC approach in M9 minimal medium. Unlabeled L-lysine (Lys-0) is used for precultures. At exponential phase, cells are washed and pulse-labeled with Lys-8 in the presence of ampicillin (+ Amp) for 4 h. A Lys-8-treated culture without ampicillin is used as labeling control. Proteins are analyzed by LC-MS/MS. Heavy (H) to light (L) protein ratios are calculated to assess Lys-8 incorporation. (b, c) Histograms of Lys-8 incorporation (%). The number (n) of analyzed proteins, the average Lys-8 incorporation, and the coefficient of variation (CV) are indicated. (d) Significance B analysis of Lys-8 incorporation in a representative experiment. Biological triplicates were used to identify proteins with a significantly increased biosynthesis in persister cells (blue dots)

Analyzing of Persister Proteomes 151

152

Bork A. Berghoff

Fig. 2 Proteome analysis of postantibiotic recovery using LC-MS/MS and label-free quantification. (a) Top: Illustration of a recovery experiment. Cultures are treated with ampicillin (+ Amp) to lyse growing cells (blue) and to isolate persister cells (orange). Persister cells are washed and transferred to fresh medium to allow recovery (purple). Bottom: Representative growth curve of a recovery experiment in LB medium. An ampicillintreated sample (amp) and recovery samples (rec1 and rec1.5) are withdrawn at the indicated time points, followed by LC-MS/MS analysis and label-free quantification. (b, c) Volcano plots illustrating p-values ( log10) versus protein ratios (log2) as determined by label-free quantification. The number (n) of quantified proteins is indicated. Proteins with significant changes during the recovery phase are highlighted as red dots

transferred to fresh medium to allow awakening and recovery of persister cells (Fig. 2a). Persister proteomes were compared at different time points during the recovery phase using LC-MS/ MS and label-free quantification (Fig. 2b, c), which identified proteins with importance to the recovery process [11]. We propose that the methods presented here will be useful to study persister proteomes in many model bacteria. Accumulation of proteome data from different bacterial persister fractions, and their comparison, might reveal universal persistence factors. This in turn will provide means for tackling persister cells in clinical settings.

2 2.1

Materials Bacterial Strains

2.2 Media and Reagents

The methods described here were applied to study the persister proteome of the high persistence mutant Δ1–41 ΔistR [11]. This mutant is derived from E. coli K-12 wild type MG1655 and has up to 100-fold more persister cells during exponential phase than the wild type when treated with the beta-lactam antibiotic ampicillin or the fluoroquinolone antibiotic ciprofloxacin [15]. High persister levels (>1%) are beneficial for proteome analysis, since they simplify the persister isolation step and enable deep proteome profiling. For SILAC, minimal media are preferable (see Note 1). For labelfree methods, any medium composition might be used. Here, we used M9 minimal medium and lysogeny broth (LB) for cultivation of E. coli.

Analyzing of Persister Proteomes

153

1. LB medium (Miller): Mix 10 g tryptone, 5 g yeast extract, 10 g NaCl, and 1 L distilled water and autoclave. 2. M9 salts (5 stock solution): Dissolve 64 g Na2HPO4·7H2O, 15 g KH2PO4, 2.5 g NaCl, and 5 g NH4Cl in 1 L distilled water and autoclave. 3. 1 M MgSO4 solution: Dissolve 24.65 g MgSO4·7H2O in 100 mL distilled water and autoclave. 4. 1 M CaCl2 solution: Dissolve 11.1 g CaCl2 in 100 mL distilled water and autoclave. 5. 1 mg/mL thiamine solution: Dissolve 10 mg thiamine hydrochloride in 10 mL distilled water. Sterilize by filtration (0.22 μm), prepare 1 mL aliquots and store at 20  C. 6. 20% (w/v) glucose solution: Dissolve 20 g glucose in 100 mL distilled water and autoclave. 7. M9 minimal medium: Mix 777 mL water, 200 mL of M9 salts (5), 20 mL of 20% (w/v) glucose, 2 mL of 1 M MgSO4, 1 mL of 1 mg/mL thiamine, and 100 μL of 1 M CaCl2. CaCl2 should be added last to avoid precipitation. 8. 10 mg/mL L-lysine solution: Dissolve 10 mg regular L-lysine (Lys-0) or L-lysine-13C6,15N2 (Lys-8; see Note 2) in 1 mL distilled water, sterilize by filtration (0.22 μm) and store at 4  C. 9. 100 mg/mL ampicillin solution: Dissolve 1 g ampicillin sodium salt in 10 mL distilled water. Sterilize by filtration (0.22 μm), prepare 1 mL aliquots and store at 20  C (see Note 3). 10. 0.9% (w/v) NaCl solution: Dissolve 9 g NaCl in 1 L distilled water and autoclave. 2.2.1 Protein Sample Preparation

1. Sodium dodecyl sulfate (SDS) lysis buffer: Dissolve 4 g SDS in 90 mL distilled water and add 10 mL of 1 M Tris–HCl (pH 7.6). Store at room temperature. 2. Urea buffer: Dissolve 36 g urea and 15.2 g thiourea in 90 mL distilled water and add 1 mL of 1 M HEPES buffer. Check the pH and, if necessary, adjust to a final pH of 8. Fill up with distilled water to reach a final volume of 100 mL. Prepare 10 mL aliquots and store at 20  C. 3. 1 M dithiothreitol solution: Dissolve 1.54 g dithiothreitol in 10 mL distilled water. Prepare 1 mL aliquots and store at 20  C. 4. 550 mM iodoacetamide solution: Dissolve 101.7 mg iodoacetamide in 1 mL distilled water. Prepare fresh and store in the dark until use.

154

Bork A. Berghoff

5. 0.5 μg/μL Lys-C: Add the desired amount of distilled water to the vial supplied by a company. For example, add 100 μL distilled water to 50 μg Lys-C. Prepare 10 μL aliquots and store at 20  C. 6. 0.5 μg/μL trypsin: Add the desired amount of distilled water to the vial supplied by a company. For example, add 100 μL distilled water to 50 μg trypsin. Prepare 10 μL aliquots and store at 20  C.

3

Methods

3.1 Pulsed-SILAC of Persister Cells

1. Add 5 mL of M9 minimal medium to a 50-mL Erlenmeyer flask and add 15 μL of 10 mg/mL Lys-0. The final concentration of L-lysine is 30 μg/mL. Inoculate a single colony and grow the culture overnight (~16 h) at 37  C with shaking (180 rpm). Cells will adapt to L-lysine utilization due to the presence of Lys-0 during overnight incubation. Use appropriate selection markers depending on the strain. 2. The next day, add 10 mL of M9 minimal medium to a 100-mL Erlenmeyer flask, add 30 μL of 10 mg/mL Lys-0 and 100 μL overnight culture (dilution of 1:100). Grow the culture at 37  C with shaking (180 rpm) until exponential phase (OD600 of 0.3–0.4) is reached. 3. Transfer the culture to a centrifugation tube and pellet cells by centrifugation (10,000  g, 3 min, room temperature). Resuspend the cell pellet in 1 mL of M9 minimal medium (without L-lysine) and transfer to a 1.5-mL reaction tube. After centrifugation in a microcentrifuge (10,000  g, 3 min, room temperature), resuspend the cell pellet in 1 mL of M9 minimal medium (without L-lysine). These washing steps will remove Lys-0. 4. Add 1 mL of resuspended cells to a 100-mL Erlenmeyer flask, containing 9 mL of M9 minimal medium, 30 μL of 10 mg/mL Lys-8 and 20 μL of 100 mg/mL ampicillin. Incubate at 37  C with shaking (180 rpm) for 4 h (see Note 4). 5. As a control, prepare and incubate a culture as described in step 4, but omit ampicillin. 6. Transfer the cultures from steps 4 and 5 to a centrifugation tube and pellet cells by centrifugation (10,000  g, 3 min, 4  C; see Note 5). Resuspend each cell pellet in 1 mL of ice-cold M9 minimal medium (without L-lysine) and transfer to a 1.5-mL reaction tube. After centrifugation in a microcentrifuge (10,000  g, 3 min, 4  C), cell pellets can be stored at 20  C until sample preparation for MS analysis.

Analyzing of Persister Proteomes

3.2 Persister Recovery

155

1. Add 5 mL of LB medium to a 50-mL Erlenmeyer flask, inoculate a single colony and grow the culture overnight (~16 h) at 37  C with shaking (180 rpm). Use appropriate selection markers depending on the strain. 2. The next day, add 20 mL of LB medium to a 200-mL Erlenmeyer flask and add 200 μL of overnight culture (dilution of 1:100). Grow culture at 37  C with shaking (180 rpm) until exponential phase (OD600 of 0.4–0.5) is reached. 3. Add 40 μL of 100 mg/mL ampicillin and incubate at 37  C with shaking (180 rpm) for 2 h (see Note 4). 4. For ampicillin-treatment samples (see Fig. 2), transfer the culture from step 3 to a centrifugation tube and pellet the cells by centrifugation (10,000  g, 3 min, 4  C; see Note 5). Resuspend the cell pellet in 1 mL of ice-cold 0.9% NaCl and transfer to a 1.5-mL reaction tube. After centrifugation in a microcentrifuge (10,000  g, 3 min, 4  C), cell pellets can be stored at 20  C until sample preparation for MS analysis. 5. For recovery samples (see Fig. 2), transfer the culture from step 3 to a centrifugation tube and pellet cells by centrifugation (10,000  g, 3 min, room temperature). Resuspend the cell pellet in 1 mL of 0.9% NaCl and transfer to a 1.5-mL reaction tube. After centrifugation in a microcentrifuge (10,000  g, 3 min, room temperature), resuspend the cell pellet in 1 mL fresh LB medium and add to a 200-mL Erlenmeyer flask, containing 19 mL LB medium. 6. Incubate the culture from step 5 at 37  C with shaking (180 rpm) to allow recovery of persister cells. After an appropriate time (see Note 6), harvest the cells as explained in step 4. Cell pellets can be stored at 20  C until sample preparation for MS analysis.

3.3 Sample Preparation for MS

1. Resuspend the cell pellet in 20–40 μL of SDS lysis buffer (see Note 7). Incubate for 10 min at 70  C and sonicate using a sonotrode (10 pulses at 20% output intensity for 1 s). This procedure will lyse the cells and homogenize the sample. 2. Remove cell debris by centrifugation in a microcentrifuge (16,000  g, 10 min, room temperature; see Note 8). Solubilized proteins are now in the supernatant. 3. Transfer the supernatant to a 1.5-mL reaction tube and determine the protein concentration (see Note 9). 4. Transfer ~10 μg protein to a new 1.5-mL reaction tube (see Note 10). 5. Add 4 volumes of acetone (100%) and incubate at 20  C for 1 h to precipitate proteins. Centrifuge in a microcentrifuge (14,000  g, 10 min, 4  C) to pellet proteins. Add 100 μL

156

Bork A. Berghoff

acetone (90%), vortex gently and repeat the centrifugation step. Open the lid of the reaction tube and dry the protein pellet at room temperature (e.g., in a fume hood; see Note 11). 6. Add 20 μL urea buffer to the protein pellet and resolve by gentle pipetting (see Note 12). 7. Add 10 mM dithiothreitol and incubate at room temperature for 30 min. Then, add 55 mM iodoacetamide and incubate at room temperature for 20 min in the dark. These steps will reduce disulfide bonds and alkylate sulfhydryl groups of cysteine residues. 8. Enzymatic cleavage of proteins (in-solution digest) is performed in a two-step protocol. First, Lys-C is added in a 100:1 protein-to-enzyme ratio, and reactions are incubated for 3 h at room temperature. Next, trypsin is added in a 100:1 protein-to-enzyme ratio, and reactions are incubated overnight at room temperature. 9. The resulting peptides are desalted using stop and go extraction (STAGE) tips (see Note 13). 3.4

4

MS Analysis

Detailed information on how LC-MS/MS analysis is performed would exceed the scope of this chapter. For analysis of the high persistence mutant Δ1–41 ΔistR, an UHPLC system (EASY-nLC 1000, Thermo Fisher Scientific) and a QExactive HF Orbitrap mass spectrometer (Thermo Fisher Scientific) were applied [11]. We would like to refer the reader to a detailed description of technical specifications for LC-MS/MS [22]. For processing of MS raw data, we recommend the MaxQuant software package and the implemented Andromeda search engine [21]. For statistical analysis, including Significance B analysis, we recommend the Perseus software platform [23].

Notes 1. For SILAC experiments, minimal media should not contain the specific amino acid that is used for metabolic labeling of proteins. Unlabeled amino acids in the medium will compete with stable isotope amino acids for incorporation into proteins, which will impair downstream analysis and quantification. In general, there is no need for using strains that are auxotrophic for the respective amino acid. If minimal media are used, incorporation rates of stable isotope amino acids are satisfying also in prototrophic strains [11, 17, 18]. 2. Here, we used L-lysine-13C6,15N2 (Lys-8) for metabolic labeling of proteins. However, several alternatives exist, including Llysine-13C6 (Lys-6) and L-arginine-13C6 (Arg-6). Furthermore, combinations of L-lysine and L-arginine isotopes can be

Analyzing of Persister Proteomes

157

applied. Some MS facilities might prefer distinct isotopes for downstream quantification processes, and we therefore strongly recommend consulting the cooperating MS facility before deciding on a stable-isotope amino acid. 3. Stability of ampicillin depends on temperature and pH. For long-term storage, we recommend to freeze aliquots at 20  C. Avoid extensive freeze–thaw cycles. Ideally, use a fresh aliquot for every experiments. 4. We used a final ampicillin concentration of 200 μg/mL, which is approximately 20 the minimum inhibitory concentration (MIC) for E. coli MG1655. Note that ampicillin at this concentration efficiently lyses growing cells, while nongrowing cells (e.g., persister cells) are largely recalcitrant. In stationary phase, almost all cells are in a nongrowing state and tolerate ampicillin. Hence, major cell lysis will not be observed in stationary-phase cultures. By contrast, in exponential-phase cultures, the effectiveness of ampicillin can be easily recognized by a significant drop in optical density (see Fig. 2a), since the majority of cells is in a growing state. Clearance of growing cells depends on the doubling time, which in turn depends on the medium composition. Hence, the ampicillin treatment duration needs to be adapted to the growth medium. In LB medium, the doubling time for E. coli MG1655 is 25–30 min, and major lysis is observed after 30 min. In M9 minimal medium, the doubling time is approximately 60 min, and major lysis is observed after 60 min. We recommend to use an ampicillin treatment duration that is four times longer than the doubling time to ensure complete lysis of nonpersisters. Note that other beta-lactam antibiotics may be used as well. For example, if very long treatment durations (i.e., several days) are mandatory, ampicillin can be replaced by carbenicillin, which is more stable than ampicillin. 5. Even though long ampicillin treatment will efficiently lyse all nonpersisters and clear the lysate [3], persister isolates after centrifugation at 10,000  g might suffer from contamination with cell debris. In case contamination with cell debris is a major problem, further purification steps (e.g., filtration) should be applied. 6. In order to determine the time points for recovery samples, growth curves should be consulted. For recovery experiments, persister cells are isolated using ampicillin treatment, followed by washing steps and transfer to fresh medium. This procedure will result in a typical recovery phase, in which no bulk growth resumption is detectable by optical density measurements (see Fig. 2a). Our results indicate that proteome analysis of samples, withdrawn at the end of the recovery phase, provides the best information on protein regulation in recovering persister cells (compare Fig. 2b, c).

158

Bork A. Berghoff

7. The volume of the lysis buffer is adjusted to the number of isolated persister cells. We used 20 μL lysis buffer for ~4  107 persister cells. If a regular sonotrode is used for sonication, the minimum volume is 20 μL lysis buffer. 8. The centrifugation temperature should not fall below 18  C to avoid precipitation of SDS. Precipitation of SDS might result in renaturation of proteins and reactivation of proteases. 9. We used the DC protein assay (Bio-Rad) for samples with a protein concentration of >0.5 μg/μL. If the protein concentration is Save Workspace As. . .). All intermediary variables and tables are stored in this file and can be retrieved quickly (Session > Load Workspace. . .) if you need to repeat the script execution. 16. The “orthogonal” variant of the classic plsr approach exploits different parts of the covariance for each linear subregression (i.e., component). In other words, the information used in the first component is not used in subsequent components. In order to avoid issues with overfitting, we generally limit the number of components to two per oplsr model. 17. Performances of models are assessed by leave-one-out crossvalidation: For a given model and for each sample, a submodel excluding the given sample is computed and the phenotype of this sample is predicted. The average of differences between measured and predicted phenotypes are then used as a performance score (i.e., Root Mean Square Error of Prediction, RMSEP). 18. The requirement of additional components to reach a local minimum of the overall error (Fig. 1a) indicates that the phenotype studied is complex and implicates different molecular mechanisms. For example, proteomic models aimed at describing antibiotic resistance were unable to accurately predict MIC values with a low number of components. Besides, efficient oplsr models will be more challenging to obtain but also more biologically meaningful with increasing strain and sample numbers.

174

Pablo Manfredi et al.

19. All graphs can be exported from RStudio as pictures by using the “Export” function on the image explorer panel. The arrow buttons on the graphics panel allow glancing over all graphs generated. In case the panel is overloaded, you can clear everything and restart the analysis using the broom icon. 20. To optimize the biological insight of persister physiology, iterative cycles of feature reduction are executed. An initial model is computed from the entire dataset and variables with a VIP above 1 are considered for the computation of the next cycle model. After each cycle, new VIP scores are computed from the new models and a variable selection is operated for VIPs above 1 for integration into the next cycle. This process greatly enhances the model predictions and best cross-validations are achieved with just few features (here: N ¼ 35).

Acknowledgments This study was supported by the Swiss National Science Foundation NRP72 grant 407240_167080 and by the Swiss National Science Foundation grant 310030_189253 to U.J. References 1. Andersson DI, Balaban NQ, Baquero F et al (2020) Antibiotic resistance: turning evolutionary principles into clinical reality. FEMS Microbiol Rev 44:171–188 2. Fauvart M, De Groote VN, Michiels J (2011) Role of persister cells in chronic infections: clinical relevance and perspectives on antipersister therapies. J Med Microbiol 60:699–709 3. Cohen NR, Lobritz MA, Collins JJ (2013) Microbial persistence and the road to drug resistance. Cell Host Microbe 13:632–642 4. Hansen CR, Pressler T, Hoiby N (2008) Early aggressive eradication therapy for intermittent Pseudomonas aeruginosa airway colonization in cystic fibrosis patients: 15 years experience. J Cyst Fibros 7:523–530 5. Santi I, Manfredi P, Maffei E, Egli A, Jenal U (2020) Evolution of antibiotic tolerance shapes resistance development in chronic Pseudomonas aeruginosa infections. mBio 12: e03482–e03520 6. Fridman O, Goldberg A, Ronin I et al (2014) Optimization of lag time underlies antibiotic tolerance in evolved bacterial populations. Nature 513:418–421

7. Levin BR, Concepcion-Acevedo J, Udekwu KI (2014) Persistence: a copacetic and parsimonious hypothesis for the existence of non-inherited resistance to antibiotics. Curr Opin Microbiol 21:18–21 8. Thevenot EA, Roux A, Xu Y et al (2015) Analysis of the human adult urinary metabolome variations with age, body mass index, and gender by implementing a comprehensive workflow for univariate and OPLS statistical analyses. J Proteome Res 14:3322–3335 9. Kreeger PK (2013) Using partial least squares regression to analyze cellular response data. Sci Signal 6:tr7 10. Mevik BH, Wehrens R, Liland KH, Hiemstra P (2020) R package pls: partial least squares and principal component regression. https:// CRAN.R-project.org/package¼pls. Accessed 19 Nov 2020 11. Martens H (2001) Reliable and relevant modelling of real world data: a personal account ofthe development of PLS regression. Chemom Intell Lab Syst 58:85–95 12. Ahrne E, Glatter T, Vigano C et al (2016) Evaluation and improvement of quantification accuracy in isobaric mass tag-based protein

Proteomic Signatures of Anitmicrobial Persistence quantification experiments. J Proteome Res 15:2537–2547 13. Janes KA, Albeck JG, Gaudet S et al (2005) A systems model of signaling identifies a molecular basis set for cytokine-induced apoptosis. Science 310:1646–1653 14. Kumar D, Srikanth R, Ahlfors H et al (2007) Capturing cell-fate decisions from the molecular signatures of a receptor-dependent signaling response. Mol Syst Biol 3:150 15. Kreeger PK, Mandhana R, Alford SK et al (2009) RAS mutations affect tumor necrosis factor-induced apoptosis in colon carcinoma cells via ERK-modulatory negative and positive

175

feedback circuits along with non-ERK pathway effects. Cancer Res 69:8191–8199 16. Wall L, Christiansen T, Orwant J (2000) Programming perl, 3rd edn. O’ Reilly Media, Sebastopol 17. R Core Team (2019). R: a language and environment for statistical computing. R Foundation for Statistical Computing. https://www. R-project.org/. Accessed 19 Nov 2020 18. Chong I-G, Jun C-H (2005) Performance of some variable selection methods when multicollinearity is present. Chemom Intell Lab Syst 78:103–112

Chapter 12 The Use of Experimental Evolution to Study the Response of Pseudomonas aeruginosa to Single or Double Antibiotic Treatment Isabella Santi, Pablo Manfredi , and Urs Jenal Abstract The widespread use of antibiotics promotes the evolution and dissemination of drug resistance and tolerance. Both mechanisms promote survival during antibiotic exposure and their role and development can be studied in vitro with different assays to document the gradual adaptation through the selective enrichment of resistant or tolerant mutant variants. Here, we describe the use of experimental evolution in combination with time-resolved genome analysis as a powerful tool to study the interaction of antibiotic tolerance and resistance in the human pathogen Pseudomonas aeruginosa. This method guides the identification of components involved in alleviating antibiotic stress and helps to unravel specific molecular pathways leading to drug tolerance or resistance. We discuss the influence of single or double drug treatment regimens and environmental aspects on the evolution of antibiotic resilience mechanisms. Key words Experimental evolution, Antibiotic tolerance, Adaptation, Cyclical treatment, Pseudomonas aeruginosa, Aminoglycosides, Fluoroquinolones

1

Introduction The great therapeutic achievements of antibiotics for human health have been dramatically undercut by the steady evolution of survival strategies allowing bacteria to overcome antibiotic action [1, 2]. While resistance plays a major role in antibiotic-treatment failure, bacteria can use other mechanisms such as tolerance to survive antibiotic treatment [3]. Resistance is generally drug specific and leads to an increase of the minimum inhibitory concentration (MIC). Tolerance is the ability of antibiotic-sensitive bacteria to survive during transient exposure to different antibiotics at concentrations that exceed their MIC for a particular drug [4]. Experimental evolution has proven an effective tool to probe the gradual accumulation of antibiotic resistance [5–9] and tolerance [10–15] as well as to study how drug tolerance can influence the

Natalie Verstraeten and Jan Michiels (eds.), Bacterial Persistence: Methods and Protocols, Methods in Molecular Biology, vol. 2357, https://doi.org/10.1007/978-1-0716-1621-5_12, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2021

177

178

Isabella Santi et al.

development of drug resistance [16–19]. Here, we introduce an experimental evolution protocol that implements recurrent population genome sequencing to investigate gradual genetic adaptations of evolving populations. This allows following the gradual evolution of sensitive ancestors of P. aeruginosa into different competing sublineages carrying single or multiple mutations conferring tolerance or resistance and its dependence on the specific drug treatment applied. By evolving multiple lineages in parallel, this approach provides a comprehensive view of the adaptation processes and of the genetic landscape behind these processes. This experimental evolution protocol uncovers important tolerance determinants and reveals the role of low-level resistance mutations in the promotion of P. aeruginosa survival at high therapeutic antibiotic doses. This is of particular importance, as it allows to dissect the specific contribution of bona fide tolerance mutations and resistance to the overall survival of this important human pathogen during specific pharmacokinetic regimens [20].

2

Materials

2.1 General Materials

Prepare all solutions and media using deionized water and store at room temperature, unless stated otherwise. 1. Antibiotic stock solution of 10 mg/mL tobramycin and 2 mg/ mL ciprofloxacin: Dissolve 100 mg of tobramycin powder stored at 4
C in 10 mL of sterile ultrapure water or dissolve 5 mg of ciprofloxacin powder stored at 4
C in 2.5 mL of sterile ultrapure water. Filter-sterilize (0.22μm) and store aliquots immediately at 20
C (see Note 1). 2. Lysogeny Broth (LB) medium: Dissolve 10 g tryptone, 5 g yeast extract, and 10 g NaCl in 1 L deionized water, and autoclave for 30 min at 121
C (see Note 2). 3. LB agar: Weigh 10 g tryptone, 5 g yeast extract, 10 g NaCl and 15 g agar and add water to a final volume of 1 L before autoclaving. Allow the medium to cool to 50–60
C before pouring in Petri dishes. Store plates at 4
C no longer than 1 month. 4. Sterile glass test tubes and Erlenmeyer flasks of 100 mL (see Note 3). 5. Sterile plastic tubes of 15 and 50 mL, suitable for centrifugation. 6. Spectrophotometer. 7. Benchtop microcentrifuge capable of 4000  g. 8. Centrifuge for 50 mL tubes capable of 7000  g. 9. PCR thermocycler.

Time-Resolved Evolution of Bacterial Drug Response

179

10. Incubator at 37
C capable of shaking at 170 rpm. 11. Multichannel pipette for plating. 2.2 In Vitro Evolution Experiment

1. LB-DMSO cryoprotectant—LB medium with 15% (v/v) DMSO: Mix 15 parts of DMSO with 75 parts of LB. 2. Polypropylene tubes suitable for 80 (cryotubes).

2.3 Determination of MIC




C preservation

1. Sterile 96-well microplates with lid and breathable sterile adhesive seals. 2. Plate reader (OD600 nm).

2.4 Genomic DNA Preparation

1. GenEluteBacterial Genomic DNA kit (Sigma) or equivalent. 2. Nanodrop spectrophotometer. 3. 10 TBE buffer: Dissolve 108 g Tris base and 55 g boric acid in 900 mL of water and 40 mL of 0.5 M EDTA solution (pH 8.0). Adjust the volume to 1 L before autoclaving. 4. 0.6% agar gel: Weigh 0.6 g of agarose, add 100 mL of TBE buffer before melting in the microwave.

2.5 Isolation of Hypertolerant or Drug Resistant Strains

3

1. Sterile 96-well microplates with lid and breathable sterile adhesive seals.

Methods Perform all handlings at room temperature and work under sterile conditions. Incubation is carried out at 37
C, shaking at 170 rpm for liquid cultures. Centrifugation steps are performed at 4000  g in the microcentrifuge and at 7000  g in the larger floor model centrifuge.

3.1 Evolution Experiment

Evolution experiments often include multiple consecutive days of lab work. Also, many factors can potentially confound the outcome of the experiment, including contaminations, timing of growth and treatment periods, accidental events like power breakdown or glassware breakage. To improve the reproducibility of the experiments and increase work efficiency, careful material preparations, thoughtful experimental set-up and careful handling of evolving populations are of critical importance. It is also important to note that protocols and drug concentrations used were optimized for P. aeruginosa (see Note 4).

180

Isabella Santi et al.

To start the evolution experiment: 1. Streak the ancestor strain from LB-DMSO stock (see Note 4) on a fresh LB plate just before the start of the evolution experiment. 2. Inoculate single colonies into individual test tubes containing 5 mL of LB and incubate overnight. One colony is used as the founder of each evolving lineage (see Note 5). 3. Measure the optical density of the overnight cultures at 600 nm (OD600) and dilute to an initial OD600 of 0.12 (corresponding to ~108 CFU/mL) in Erlenmeyer flasks containing 20 mL of LB. For this, centrifuge the corresponding volume of bacteria in a 1.5 mL Eppendorf tube, remove the supernatant and resuspend the pellet in 20 mL of LB. Before adding the antibiotic or combination of antibiotics (see Notes 1, 6, and 7), withdraw an aliquot of bacteria (200μL is usually sufficient) to determine the total CFU per mL at time zero (T0). 4. Add antibiotic(s) and incubate the flasks on a shaker for the entire duration of the experiment (see Note 8). 5. To determine CFUs per mL at T0, make serial tenfold dilutions of the harvested 200μL aliquot, spot droplets of 10μL (see Note 9) on LB agar plates and incubate plates overnight (see Note 10). 6. Harvest an aliquot for the freezer stock. Centrifuge 1 mL of the overnight cultures resulting from step 2 for 1 min, resuspend the pellet in 1 mL of LB-DMSO and store immediately at 80
C (see Note 11). 7. Harvest an aliquot for population genome sequence analysis (see Note 11 and Subheading 3.3). Centrifuge 1 mL of the overnight cultures resulting from step 2 for 1 min, remove the supernatant and store the cell pellet immediately at 80
C. 8. Harvest an aliquot to determine the MIC (see Subheading 3.2). 9. Harvest an aliquot to determine CFUs. Make appropriate serial tenfold dilutions of the samples in LB medium, spot droplets on LB agar plates and incubate up to 48 h to quantify surviving cells (see Notes 9, 10 and 12). The fraction of survivors is calculated by dividing the number of CFU/mL after 3 h treatment by the number of CFU at T0 (before antibiotic treatment). 10. Transfer the remaining culture to a 50 mL tube using a sterile pipette to minimize cross-contamination (see Note 13), centrifuge for 10 min and wash twice with LB to remove the antibiotic (see Note 12). Resuspend the bacterial pellet in a glass tube containing 5 mL of LB and incubate overnight to start the next growth cycle (Fig. 1) (see Note 7).

Time-Resolved Evolution of Bacterial Drug Response

181

Fig. 1 Schematic of the experimental design for the iterative exposure of P. aeruginosa to bactericidal antibiotics

11. Go to step 3 and repeat the procedure with daily cycles until drug resistance or tolerance has evolved (Fig. 1). Under these conditions, evolution of drug resistance is generally observed within 6–7 days for single drug treatment and low level resistance is observed within 8–9 days for double drugs treatment (see Note 14). 12. For data analysis, plot the fraction of survivors (step 9) and MIC values (see Subheading 3.2) as a function of time. Include the fraction of cells carrying specific single-nucleotide polymorphisms (SNPs) as pie-chart for daily samples (see Subheading 3.3). By directly linking the increase in resistance and bacterial survival with genetic changes occurring during evolution, these data allow discriminating mutations leading to resistance and tolerance. To establish causal relationships between genetic changes and antibiotic phenotypes, individual mutations need to be further characterization as describe in Subheadings 3.4 and 3.5 (see Note 15). 3.2 MIC Determination

MIC values are determined for the ancestor before the start of the evolution experiment and periodically for evolving populations and isolated clones during the experiment. Subheading 3.2 is an adaptation from previously described methods [21]. 1. Dilute the overnight culture (step 3 of Subheading 3.1) into LB medium to reach an OD600 of 0.1. 2. Make twofold serial dilutions of the antibiotic in a 96-well plate. To do so, add 200μL of LB to column 2 and supplement it with the antibiotic to reach the double of the highest concentration to be tested. Add 100μL of LB to columns 3–10 and transfer 100μL from each column to the next, starting with column 2. Add 100μL of LB to column 11 as a control lacking

182

Isabella Santi et al.

antibiotics. Add 100μL of the bacterial solution prepared in step 1 to each well and fill the wells at the edges with LB as a sterility control. Cover the microplate with a breathable seal and plastic lid, and incubate for 16–20 h. 3. Measure the OD600 nm for each well in a microplate reader. Verify the positive and negative controls for adequate microbial growth and medium sterility. The MIC represents the lowest concentration with clear growth inhibition. 3.3 Preparation of gDNA and Genome Sequencing

Genomic DNA (gDNA) is extracted from the pellets of frozen overnight cultures (step 7 in Subheading 3.1) of the evolution experiment. 1. Perform gDNA preparation following the manufacturer’s instructions. We used the GenEluteBacterial Genomic DNA kit (Sigma) with exception of the final elution step, for which we used H2O instead of the elution buffer provided with the kit. 2. Quantify the gDNA concentration with a NanoDrop spectrophotometer. 3. Monitor genomic DNA integrity by running an aliquot on a 0.6% agar gel. 4. Analyze gDNA by Illumina sequencing. Generate the library and the sample barcoding with the NexteraXT approach (Illumina) and verify the library quality with a Fragment Analyzer (Advanced Analytical). Perform PE125 sequencing runs on parallel libraries on a HiSeqlane (Illumina) with a targeted coverage of about 100. 5. Analyze DNA sequences. Map the sequencing reads onto the genome of the reference strain P. aeruginosa PAO1 (NC_002516) (see Note 5) with Bowtie2 [22] and small polymorphisms. Spot the structural and coverage variants with Samtools [23] and a collection of in-house Perl scripts [24] (see Note 16). Details for this analysis are provided in Subheadings 3.3.1–3.3.3.

3.3.1 Alignment to a Reference Genome

In order to map the reads (see Note 17) onto a reference genome, we recommend to use the alignment freeware Bowtie2 that can be installed on Windows, MacOS or Linux OS (https://sourceforge. net/projects/bowtie-bio/files/bowtie2/) [22]. Bowtie2 is command line software that requires the command terminal available on any given system. First, build index files from the desired reference. Together with the reads file (.fastq), the index files are used by Bowtie2 to perform the alignments against the target reference. This requires the sequence “.fasta” file of the targeted chromosome (for example, P. aeruginosa PAO1 is NC_002516.2, available at NCBI).

Time-Resolved Evolution of Bacterial Drug Response

183

> bowtie2-build reference.fasta reference.index

Reads from a given evolution time point (e.g., A) are mapped as follows. > bowtie2 --local -a -R 3 -N 0 -L 20 -i S,1,0.50 -x ref.index -1 A.PER.1.fastq.gz -2 A.PER.2.fastq.gz -S A.Assem.sam

The alignment is performed locally (--local, i.e., partial matches allowed), all matches are reported (-a), with high sensitivity parameters (-R 3 -N 0 -L 20 -i S,1,0.50). Here, two pair end read files are mapped (-1 PER.1.fastq.gz -2 PER.2.fastq. gz). If only a single read file is available, use -U Reads.fastq. gz instead (see Note 18). # create a binary version of the assembly file > samtools view -F 0x0100 -b A.Assem.sam > A.Assem.bam # Order the file along the reference sequence > samtools sort A.Assem.bam -o A.Assem.sort.bam # Compile the assembly in parser readable pileup format > samtools mpileup -A -B -x -d 1000 -f ref.fasta A.Assem.sort. bam > A.Assem.sort.pileup

The resulting assembly file in “.sam” or “.bam” format (binary equivalent) contains the alignment information required for SNP calling and for the identification of recombination events for a single sample (i.e., evolution time point). In order to reformat the assembly file before the identification of possible mutations (i.e., variant calling), use the Samtools and BCFtools freeware that is available under Linux (http://www.htslib.org/download/) [23]. If you do not have access to a Linux environment (e.g., https://ubuntu.com/), the software can be emulated in MacOS or windows using a virtual machine solution (e.g., https://www. virtualbox.org/).

base

T

A

A

A

G

A

G

A

C

C

G

Reference

AE004091.2_PAO1 3

AE004091.2_PAO1 4

AE004091.2_PAO1 5

AE004091.2_PAO1 6

AE004091.2_PAO1 7

AE004091.2_PAO1 8

AE004091.2_PAO1 9

AE004091.2_PAO1 10

AE004091.2_PAO1 11

AE004091.2_PAO1 12

AE004091.2_PAO1 13

#pileup file

43

43

43

43

43

43

43

43

43

43

43

coverage

match

....$....$...,,,$,,,,$,,,,,,,$,,,,$,,

............,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

............,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

............,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

............,,+3ATG,,,,,,,,,,,,,,,,,

............,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

............,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

............,,,,,,,,,A,,,,,,,,,a,,,,,,,,,,

............,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

.^...^.^....,,,^,,,^,,,,^,,,,,,,,,,,,,,,

match quality (ASCII coded)

G9CGGGGGGGG>FFGG